The XRay.Tech Blog

Are you spending too much on Zapier or Make? Meet n8n, the powerful automation provider that lets you build workflows and AI agents for a fraction of the cost. 

Whether you need enterprise-grade security with self-hosting or just want to save money on your automated setup, n8n has you covered.

In this beginner's guide, we'll walk you through setting up n8n and creating your first automation from scratch - a simple workflow that sends Slack notifications when you receive specific emails.

What Is n8n?

n8n is a flexible automation platform that connects your apps and services, allowing you to build custom workflows without coding knowledge. What sets it apart is its competitive pricing and the option for self-hosting, giving you more control over your data and potentially significant cost savings compared to other automation platforms.

Additionally, if you or your team have a background in software development, you’ll find that n8n offers extensive support for inserting custom code into your automated workflows. However, in this guide, we’ll focus on n8n’s no-code automation capabilities. 

Getting started with n8n

When you first log into n8n, you'll see the “Overview” page which will give you a summary of your workflows, credentials, and executions. For new users, this will be empty and inviting you to create your first automation. 

Simply click the "Create" button in the top right corner to create a new automated workflow.  

‍

‍

Building your first workflow: Gmail to Slack notification

Let's create something practical and immediately useful - a workflow that notifies you in Slack when you receive important emails. We'll start simple and then add some intelligence to the system.

Our workflow will:

• Detect new emails in your Gmail inbox
• Send details to you in Slack with a direct link to the email
• Filter for emails containing a specific term in the subject ("marketing")
• Share emails matching the filter with a team channel

Bear in mind that this is just one example that we’re using to demonstrate how n8n works. These same principles will apply to the hundreds of apps supported in n8n, like Jira, Hubspot, Notion, and Wordpress. 

You can find the full list of n8n’s integrations here. 

Setting up your workflow’s trigger: New emails in Gmail

Choose a trigger type

Click "Add first step" on your canvas, and you'll see a side menu appear with various trigger options. 

‍

‍

We want our workflow to run when something happens in another app, so select "Run on an app event."

‍

Pick an app and event for your trigger

Search for your trigger app (Gmail) in the search box and select it. 

‍

‍

Then, choose the specific event that will trigger your automation. You'll only see one option here for Gmail  - "On message received" - which is exactly what we need. Select it and move to the configuration screen.

‍

‍

Connect an account to authorize your trigger app

Next, you’ll need to connect a credential for your trigger app. In this case, that means signing in to your Gmail account to authorize n8n to look for new emails in your inbox. 

Click on the credentials dropdown and create a new credential. 

‍

‍

This will prompt you to sign in with your Google account and allow n8n to access your emails. 

‍

‍

Don't worry - n8n only uses this access to detect new messages and read their contents for your automation, similar to how Zapier or other platforms work. It will only access your account how you instruct it to in your automations. 

‍

‍

Once you’ve created your Gmail credential, click on the ‘x’ in the top right corner of the credential window to return the configuration menu for your trigger. 

‍

Configure your trigger

On the configuration screen, you'll see several options. Here’s the setup we’ll use:

‍

• Poll times: Determines how often n8n checks for new emails. We’ll use the default option, and search for new emails every minute.

• Simplify: Toggle this off to receive comprehensive email data. For this tutorial, make sure to keep this setting off. 

• Filters: Optional restrictions for specific senders or labels. You won’t need to apply any filters in this tutorial, but you can add them if you’d like to narrow down the list of emails your automation runs for. 

‍

Test your trigger

Once your configuration is all set, click "Fetch test event" to retrieve a sample email. 

‍

‍

On the right hand “Output” panel, you’ll see a preview of the data your workflow will process when it triggers. 

‍

‍

Switch to the "Schema" view for a clean list of all variables like subject, sender address, and timestamp - you'll use these later when crafting your Slack notification.

‍

Creating an automated action: Direct message in Slack

With your Gmail trigger configured, it's time to decide what happens when a new email arrives. Go back to your canvas, and click the "+" button to the right of your Gmail node to add an automated action to your workflow. 

‍

‍

You’ll see several options for what happens next, including flexible data transformation and flow control modules. For this tutorial, we’ll choose "Add an action in another app" so we can set up an automated Slack message. 

‍

Choose your app and action

Search for the app you want to automate – in this case, Slack. Then, select it from the search results.

‍

‍

Next, you’ll need to pick the action that you want to automate within that app. For our example, that will be "Send a message." 

‍

Connect an account for your automated action

Just like with our Gmail trigger, you'll need to connect your Slack account by creating a new credential. 

‍

‍

You might see a warning that n8n isn't officially approved by Slack – this is normal for many automation platforms and nothing to worry about. This just means that n8n hasn’t yet gone through the formal process of creating a Slack app. 

‍

‍

After connecting your Slack account, return to the configuration screen and select it from the credentials dropdown.

Configure your automated action

Now for the fun part – designing your notification message. First, decide where to send it by selecting "User" and choosing yourself as the recipient. 

‍

‍

Tip - you can select users from a list of their display names, or by entering their Slack ID. 

‍

With your user chosen, you can craft your message with a mix of static text and dynamic content from your email. To add dynamic content to your message, just drag data from the “Inputs” schema on the left into the text field. 

‍

In our example, we’ll include dynamic data for the email’s subject, the “from” address, and the time the email was sent. Note that once you drop a variable into the text window, it will transform into a bit of JSON formatted code. 

‍

‍

We’ll also include a link to view the email in our inbox. n8n didn’t retrieve a complete URL in the Gmail trigger, but it did retrieve an ID for the email. We can use this in combination with some static text to construct a complete link. 

‍

Enter [start of link[ directly into the text field, then drag and drop the “ID” variable to complete the URL. 

‍

‍

Tip: you can click the icon in the bottom right of the message text field to expand it into a larger window. 

‍

We’ll also use some Slack-compatible markdown to embed this link into our message text. That syntax looks like this:

‍

<your link | your text >

Reformatting data with simple functions

Finally, before testing this automated action, let’s reformat the “Date” timestamp to be easier to read. We can do this with a couple of simple functions appended to the end of the $json.date object. 

‍

Use the toDateTime() function to convert the text into timestamp data, then use the format() function to reformat it however you’d like. We’ll use ‘MMM dd yyy hh:mma’ to get a result like Feb 06 2025 03:27PM. 

‍

Augment your automations with AI

Not sure how to transform your data with functions? Try asking n8n’s built-in AI assistant. Just ask it any question you have about building a workflow, and it will make its best attempt at helping you out. 

‍

In our experience, it was pretty useful, but definitely not perfect. It will get you 80% of the way there, but you’ll need to fix some of its errors and finish things up on your own. 

‍

‍

You can find the AI assistant in the bottom right hand corner while you’re building. 

Test your automated action

Once you’ve configured your message to your liking, give the action a test. Testing an action in n8n will actually perform the action, so testing a “send message” action will actually send a message in Slack. 

‍

Click on “Test step” to execute the node. 

‍

‍

After your test runs, you’ll see data appear in the “outputs” panel on the right. This is the data you can access in any future steps you add to your workflow. 

‍

You can also check Slack to see the message that got sent. 

‍

‍

After our test, we see a message complete with a correctly formatted date and a link that leads directly to the email in our inbox. 

Create a more versatile workflow with filters

Simple, straightforward automations like this can be useful in some contexts, but you’ll often need something that’s a little more flexible. In many cases, you’ll want to process emails differently based on their content and metadata. 

‍

To run different actions based on different inputs, you can use n8n’s flow control features. For example, let’s imagine you want to add a second Slack message that sends a similar alert to the Marketing channel in Slack, but only if the email’s subject contains the term ‘marketing’. 

‍

To accomplish this, you only need to add a filter after the original Slack message, then add an action for a second Slack message after the filter. 

Adding a filter

Click the "+" button after your Slack action and select “Flow” from the list of actions.

‍

‍

Then, choose "Filter" from the flow control options. 

‍

‍

This node acts as a gatekeeper, only allowing emails that match specific criteria to continue through additional steps.

Configuring your filter

Every filter consists of one or more conditions. When the automation runs, it will test the incoming data against those conditions, and will only proceed to the subsequent steps in the automation if the data matches the conditions. 

‍

For our example, we’ll just evaluate one simple condition. We’ll drag and drop the “Subject” variable from the Input window into the first value field. 

‍

‍

Tip: you may need to click “execute previous nodes” to generate input data. 

‍

Then, we’ll set the operator to "contains," and enter "marketing" as the second value. We’ll also add an option to ignore case. 

‍

‍

Now only emails with "marketing" in the subject line will pass through.

‍

Testing your filter

Once your filter is configured with all of the conditions you want, give it a test by clicking on “Test step”. 

‍

If the currently loaded test data passes the filter, you’ll see data populate the output window as normal. Otherwise, it will be sent to the “Discarded” tab. 

‍

Adding actions after a filter

With your filter set up and tested, you can add any actions you want to perform for data that passes it. Remember, these actions will not run for any data that doesn’t match your specified conditions. 

‍

We’ll just add another Slack message, but this one will go to a channel instead of a specific user. It will also include the full body of the email in question, rather than a link. 

‍

‍

You can test these actions to see their outputs as well. 

‍

Saving and Activating Your Workflow

With all your nodes configured, click "Save" in the top right corner of the builder to commit all your changes. 

‍

‍

Return to the workflows overview and make sure your new automation is toggled on. 

‍

‍

That's it! Your workflow is now live and monitoring your inbox.

From now on, you'll receive a Slack message whenever an email arrives in your inbox. If the subject contains "marketing," your team channel will also get notified with the contents.

n8n pricing overview

One of n8n's most compelling features is its pricing structure. At $50/month, the Pro plan includes 10,000 executions and unlimited users – significantly more affordable than Zapier's $170/month Team plan, which only supports 25 users and counts individual tasks rather than complete executions.

‍

‍

Self-hosting n8n

For organizations with security or compliance requirements, n8n offers a unique advantage: self-hosting. 

Available through either an enterprise plan or a free community version on GitHub, this capability gives you complete control over your automation infrastructure and data.

The community version follows n8n's "fair code" model, making the source code freely available with some restrictions on commercial use. 

While it has certain limitations (like lack of variable support), it provides an option that simply isn't available with most other automation platforms.

Affordable, self-hosted automation with less frills

n8n might lack some of the polished features found in more established platforms like Zapier, but its combination of affordability, flexibility, and self-hosting options makes it a compelling choice for individuals and teams looking to build powerful automated workflows without breaking the bank.

If you’d like to use n8n or other low-code automation platforms to create more time for your team, just reach out to XRay for a free consultation to get started. 

For more DIY automation tips and tutorials, check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

AI app builders promise to turn your ideas into functioning applications with just a few prompts. But can they really deliver? We decided to put this claim to the test by attempting to rebuild our automation search engine—a web app that originally took our engineers hundreds of hours to create—using only AI prompts through Lovable.dev.

The original XRay.Tools app

Our original app, XRay.Tools, is an automation search engine where users can search for tools they're currently using (like Slack or Zoom) and discover all the automation capabilities available across multiple platforms. 

‍

‍

After you assemble your toolbelt with a few quick searches, you can click "View Automation abilities" and see all the triggers, actions, and searches available for your apps on every automation provider.

‍

‍

This tool helps our users quickly determine where they should build their automations. It provides comprehensive information and direct links to each automation provider's documentation.

The Lovable.dev challenge

Lovable.dev claims to be your "superhuman full stack engineer" that can take you from "idea to app in seconds." Naturally, we were skeptical, but decided to give it a shot.

‍

‍

Surprisingly, in just a few hours, we were able to prototype a functional automation search engine that pulled data from a Supabase database. 

‍

‍

And we accomplished this without writing a single line of code—all interaction was through chat-based prompts with the AI.

What we were able to build

The prototype we created allows users to:

‍

• Search for the tools they want to automate (e.g., Slack and Zoom)

• See which automation providers support their software

• View a list of available triggers, actions, and searches for their tools

‍

‍

• Click through to view integration details

‍

‍

We even attempted to expand beyond our original app by including 13 different automation providers, adding newer platforms like Pipedream, Tray, Unito, Parabola, Bardeen, Relay, and Flowwise alongside established ones like Zapier and Make.

‍

The AI was also able to add fun visual touches, like a bursting confetti animation that triggers whenever a user clicks on the logo at the top of the page. You’ll also notice in the GIF below that the logo gradually changes color between different purple gradients. 

‍

‍

The final result is a visually appealing web app with basic (but limited) functionality. 

The limitations

While impressive, the prototype wasn't perfect:

• The data quality issues became apparent, with some nonsensical automation "abilities" showing up (like "notifying new Twitter followers" listed as a Slack capability)

• Technical complexity was a challenge for the AI

• The final result was more of a 60-70% solution rather than production-ready code

Where AI app builders excel

Our testing revealed that Lovable.dev particularly shines when it comes to rapidly creating visual interfaces. 

The AI excels at implementing aesthetic elements like animations, borders, and layout arrangements, and can quickly turn text descriptions into functional prototypes with remarkable accuracy.

For developers, the generated code provides a solid starting point that's often much easier to work with than manually translating a Figma design from scratch. This visual strength makes it an excellent tool for early-stage ideation and demonstration.

Where human developers still win

Despite the impressive aesthetic results, AI app builders like Lovable or Bolt.new aren't yet ready to replace human developers. 

They still struggle with technical complexity, and the critical journey from 70% to 100% completion is precisely where these tools fall short. 

Without coding knowledge to modify the AI-generated output, users will quickly hit limitations that prevent creating truly production-ready applications. The final polish, security considerations, edge case handling, and performance optimization still firmly require human expertise. 

As powerful as these AI tools are becoming, the finish line remains elusive without skilled developers to bridge the gap.

No-code tools: the best way to go to market fast

If you’re looking to build a product but don’t have an engineering team, we’d recommend using no-code tools to launch your app instead. 

‍

With tools like Softr, you can create fully functional web apps with simple visual interfaces. You might not be able to get the same visual flair that you could obtain from a few prompts with Lovable, but you’ll have precise control over your app’s functionality – and you won’t need to write any code to do it. 

‍

You can explore XRay’s case studies to see some examples of the successful no-code and low-code products we’ve built with our clients. 

‍

When you’re ready to start building, check out the tutorials on our blog or reach out to XRay for a free 15-minute consultation. 

The verdict: great for prototyping, not for production

AI app builders like Lovable.dev represent an exciting step forward in rapid prototyping. They can take you from concept to working demo faster than any previous method. However, for now, they're best used as prototyping tools rather than complete development solutions.

If you don't know how to code or modify code, you're better off sticking with traditional no-code builders despite their limitations. The key advantage is that with those tools, you'll actually reach a point where you can click "publish" on something usable.

Will this change in the future? Almost certainly. The pace of improvement in AI development tools is remarkable, and we'll be checking back in six months to see how far they've come.

For more AI and automation tips, tutorials, and news every week, subscribe to our YouTube channel and follow @XRayAutomation on social media.

Anthropic has unveiled Claude 3.7 Sonnet, its latest AI model, featuring a groundbreaking "hybrid reasoning" system. 

Unlike previous iterations that relied on separate models for different tasks, this new approach enables the model to seamlessly switch between rapid responses and deliberate, step-by-step reasoning. 

This dual capability makes Claude 3.7 not just faster but also more effective at handling complex, multi-step problems in math, physics, and coding. 

In this post, we’ll take a look at what makes 3.7 Sonnet different under the hood. Then, we’ll run some real-world tests to see how the Extended and Normal ‘Thinking Modes’ perform with a couple of math and coding challenges. 

What Makes Claude 3.7 Sonnet Different?

One of the standout features of this model is its Extended Thinking Mode, which allows for deeper, more structured reasoning when enabled. This mode is particularly useful for tackling complex mathematical proofs, intricate programming tasks, and high-stakes problem-solving. 

While 3.7 Sonnet is available to all users, the Extended thinking mode is currently only accessible for users with a pro account. 

Users can control how much processing time the model dedicates to internal deliberation by setting a "thinking budget", determining the number of tokens allocated for reflection before generating a response. 

According to benchmark tests, Claude 3.7 Sonnet significantly outperforms its predecessors in software development tasks. For example, it scored 62.3% accuracy on the SWE-bench Verified benchmark, which evaluates AI performance on real-world software issues. This makes it one of the most competent AI models for debugging and code generation today.

A New Era for AI-Powered Development

Beyond general improvements, Anthropic is also rolling out Claude Code, an experimental AI coding assistant designed to streamline programming tasks directly from the developer’s terminal. 

This marks a shift toward AI tools that actively assist in real-world coding environments rather than merely offering suggestions in chat-based interactions. Currently, Claude Code is in a limited research preview but is expected to become a valuable tool for software engineers and automation experts. 

Putting Claude 3.7 Sonnet to the test: Extended vs. Normal thinking modes

With these improvements in mind, the big question is: how well does Claude 3.7 Sonnet actually perform in real-world applications? In the next section of this post, we'll run practical tests to evaluate its ability to solve mathematical problems and write production-ready code efficiently. 

Simple math and logic

We’ll begin with a simple problem:

‍

How many times does the letter P appear in the following sentence?

Alan picked three ripe apples. 

‍

While it’s trivially easy for most people to count 4 of the letter P in that sentence, it’s the kind of basic question that frequently stumps generative AI models. 

‍

3.7 Sonnet Normal Thinking Mode

‍

And sure enough, 3.7 Sonnet failed to give a correct answer in our tests with the Normal thinking mode enabled, reporting only 3 P’s in the sentence. The double P in “apple” proved to be a point of confusion for the AI, as is often the case. 

‍

‍

3.7 Sonnet Extended Thinking Mode

‍

However, enabling the Extended thinking mode instantly provides the correct answer. 

‍

‍

As you can see in the screenshot above, the Extended thinking model makes sure to break down this simple problem into granular components. In this case, it analyzes each individual letter to assess whether or not it’s a P, and tallies up the result to give us an answer. 

‍

Of course, running the same prompt several times will likely produce different results, but this early test is a clear indication that the Extended version of 3.7 Sonnet is more reliable when it comes to math and logic. 

‍

Advanced math

Now, let’s see how both modes of the model handle a more complex problem. Here’s a question we’ve borrowed from Q-51’s GMAT practice set:

‍

If x is a positive integer such that (x - 1)(x - 3)(x - 5)....(x - 93) < 0, how many values can x take?

‍

Per Q-51, the correct answer is 23. Let’s see how both modes of 3.7 Sonnet handle this problem. 

‍

3.7 Sonnet Normal Thinking Mode

‍

When we ask 3.7 Sonnet to solve this problem using the Normal thinking mode, we can see it instantly try to break the problem down in detail to solve it. 

‍

‍

However, its final answer is not correct. It offers up 48 as the solution instead of 23. 

‍

‍

3.7 Sonnet Extended Thinking Mode

‍

When we give the exact same prompt to 3.7 Sonnet Extended, we can see that it takes much more time to try and solve the problem. 

‍

‍

It ultimately takes about 1 minute and 45 seconds to arrive at an answer, but this time it’s the correct result of 23. 

‍

‍

As we can see, ratcheting up the complexity causes 3.7 Sonnet Extended to take a much longer time in answering, but it’s worth it in the end as it’s able to provide an accurate solution. 

‍

Coding challenge

Next, let’s see how the two modes of 3.7 Sonnet handle a coding problem. We’ll ask both to make a website with a scrolling text animation using the following prompt:

‍

Create a website with HTML, CSS, and Javascript. The website is for a fictional widget company, Wetzel's Widgets. In the hero section of the site, include the text "Our Widgets can handle __". In the blank at the end of the sentence, animate a series of nouns like "Productivity", "Synergy", "Markets", etc. The animation should scroll through the list of nouns. As one noun scrolls up and disappears, the next scrolls up and takes its place.

‍

3.7 Sonnet Normal Thinking Mode

‍

Our first attempt with the normal thinking mode results in a rather well-designed web page, but no animation at all. It doesn’t even include any of the text that’s meant to be in the animated scroll. 

‍

‍

A second attempt with the Normal thinking mode performs a bit better, but the scrolling text is misaligned and floats down the page with each rotation.  

‍

‍

Unfortunately, neither of these attempts are going to be usable without significant edits. 

‍

3.7 Sonnet Extended Thinking Mode

‍

When given the same prompt, the Extended thinking mode delivers a much better result. The scrolling animation works exactly as described in the prompt. 

‍

‍

This is an excellent starting point for a website with the animation we described. 

‍

Claude 3.7 Sonnet: Finally, a Reliable AI for Code and Math

After putting Claude 3.7 Sonnet Extended through its paces, the results are clear: this version finally delivers the reliability developers and automation experts have been waiting for.

In Extended Thinking Mode, Claude consistently produced working automation scripts, accurate formulas, and stable API configurations. Whether you need to transform data, set up tool integrations, or generate custom functions for your low-code platforms, Claude 3.7 Extended delivers code that actually works the first time—a major improvement over previous iterations.

The difference was especially stark when comparing normal thinking mode with extended mode. While the default setting sometimes struggled with multi-step logic, the extended version methodically worked through complex math problems and debugged its own mistakes, making it a much more practical tool for real-world applications.

The Bottom Line

Claude is finally evolving into the reliable AI partner we've wanted for coding and mathematical tasks, saving teams hours of debugging, troubleshooting, and rewriting. If you're building automated workflows, this model can drastically reduce frustration and wasted time.

Taking AI Automation to the Next Level

Of course, even with these advancements, having the right implementation partner makes all the difference. That’s where we come in.

At XRay.Tech, we specialize in turning AI and automation into real productivity gains for your team. Instead of wrestling with repetitive tasks or spending hours fixing broken scripts, we help you:

• Identify your biggest time-wasters
• Streamline clunky processes
• Build custom low-code tools that actually fit how your team works

The result? Your team gets back to what actually matters, while AI and automation handle the rest.

Want to see how much time AI-powered automation can save your business?Book a free 15-minute consultation with a solutions engineer today.

Automating AI workflows can dramatically increase your productivity, but it comes with risks. What happens when AI generates an inappropriate response or makes a mistake? You don't want to automatically send incorrect or awkward messages to clients and colleagues. 

‍

Fortunately, you can add human oversight to your AI automations without investing in expensive enterprise software.

‍

In this guide, we'll explore two approaches to creating human-in-the-loop automations: Make's native enterprise feature and a free alternative using Airtable. Let's dive in.

‍

Option 1: Make's native “Human in the Loop” feature

‍

Make (formerly Integromat) recently introduced a dedicated human-in-the-loop module for enterprise subscribers. Here's how it works in a simple setup with two automations involving an AI prompt.

‍

1. Trigger your first automation 

Launch your first automation however you’d like. For instance, the scenario pictured below triggers when there’s a new row in a specified Google Sheet.

‍

‍

You can also add any other actions or searches you want to include to populate your 

‍

2. Generate AI content 

Send a prompt to AI tools like OpenAI or Claude to generate a response. Be sure to include any necessary data from your trigger or other previous modules in your prompt. 

‍

‍

3. Add a “Human in the Loop” module 

Add a “Human in the Loop” module, and choose “Create a review request” as the action. 

‍

‍

With this module, you can create a simple webpage hosted by Make that will include your AI response (as well as any other content you want to review). 

‍

Just add items in the “Content to review” section to add them to the page. In our example pictured below, we’ll add the AI response as an item to review. 

‍

‍

Be sure to mark the content as “editable” if you want to be able to make changes to it on the review page. 

‍

Under “Advanced settings”, you’ll see several options for styling the page. 

‍

‍

Once you’ve configured the module to your liking, click “Save”. 

‍

NOTE: while you can add this module to your Make scenarios on any plan, you’ll need an Enterprise plan to actually run it. 

‍

4. Share the “Human in the Loop” URL

Add another module to share a link to the Human in the Loop review page wherever you’d like. For example, in the automation pictured below, the link is sent to a Slack channel. 

‍

‍

5. Run the automation and approve, deny, or modify the content

When this first automation runs, it will create a page where you can review the AI-generated content, and send the link to your desired destination. 

‍

Open up the link, review the content, and approve, deny, or edit the content as appropriate. 

‍

‍

6. Trigger additional automations on approval/denial

Create a second automated scenario. For the trigger, choose “Human in the Loop”, and select “Watch completed reviews”. 

‍

‍

This trigger will launch whenever someone takes an action on a “Human in the Loop” request (e.g., Approve, Deny). 

‍

While powerful, this feature is limited to enterprise subscribers. However, there's a simple alternative that provides the same functionality without the premium price tag.

‍

Option 2: Building a Free Alternative with Airtable

‍

You can create equivalent human-in-the-loop functionality using virtually any app that allows you to easily create, reference, and edit data. 

‍

For example, you can swap out the official “Human in the Loop” step with an Airtable step instead, and you can do it with Airtable's free tier. 

‍

Airtable is a no-code database app that works great as a companion to any automation provider. You can learn more and get started with our Airtable beginner’s guide. 

‍

Alternatively, you could use the Smartuite, Google Sheets, Notion, or any number of other apps. 

‍

The process will be very similar to the setup we covered for the Human in the Loop module, and will still involve two automations (or more, if you want to run automated actions for different decisions). 

First automation: create and share a record in Airtable to review the AI output

1. Set up your trigger in Make

2. Send a request to an AI application

3. Create a new record in Airtable

After you’ve triggered your first automation and sent a request to the AI, you can store its response in an Airtable base. 

‍

You’ll want to create the database in Airtable before adding the module in Make. We’d recommend including the original request details, as well as a checkbox that you can use to indicate that the content is ready to be sent out. 

‍

‍

Finally, you should add a new view to the table. Add a filter to the view, and set it to only include records that have been checked off. We’ll use this later to trigger the second automation. 

‍

‍

Once your table is ready, just add a new Airtable module and select “Create a record” as the action. Then, map each piece of dynamic data to the appropriate Airtable field. 

‍

‍

4. Send a notification (e.g., via Slack) with the Airtable record URL

Instead of sending a Human in the Loop URL, you can send out a link directly to the relevant Airtable record. 

‍

To construct this URL, just open up any record in your Airtable base (expand the details), and copy the full URL up until the string of characters that begins with “rec”. This string is the record’s unique ID, and we’ll replace it with the dynamically retrieved ID for each newly created record. 

‍

‍

Paste this URL into your Slack message in Make (or whichever app you’re using to send the URL), and add the “ID” variable at the end. 

‍

‍

Your first automation should now be all set. You can test it out to see a Slack message with a link to the latest record in Airtable. 

‍

‍

Open it up to make sure everything looks correct, edit the AI prompt if you’d like, and check the box to get ready for the second automation. 

‍

‍

Second automation: send the approved AI output to its destination

Your second automation will trigger whenever a new record enters the designated view for approved responses. 

‍

However, you don’t necessarily need to build this automation in Make. If you’re using Airtable, and want to send the AI’s output as an email or Slack message, we’d actually recommend building it there instead. 

‍

Automations in Airtable are free and don’t limit operations in the same way that Make does. If you have a scenario in Make checking for new records even just every fifteen minutes, you’d be using up 96 operations a day with the trigger alone. This would quickly use up your monthly allotment on lower tiers.

‍

Airtable would avoid this issue, but does have limited integrations. If the app you want to send the AI output to isn’t supported in Airtable, you’ll have to use Make instead. 

‍

If you’re new to Airtable automation, you can learn how it works in the second part of our Airtable beginner’s guide.

‍

The automation that finishes up this workflow will look like this:

‍

1. Launch when a new record enters the specified view

‍

‍

2. Send the AI’s output wherever you want

In our Airtable example, we’ll send the AI output via Gmail. 

‍

‍

With the completed workflow, an AI response is generated whenever a new row is added to our Google Sheet. The response is stored in Airtable, and sent to our team for review in Slack. 

‍

We can then edit and approve the response in Airtable, and automatically send it out via Gmail. 

‍

We’ve got quick and easy oversight in our AI automation, and we don’t need an enterprise plan in Make to do it. 

Save money on AI automation with simple, DIY alternatives

Human-in-the-loop automation doesn't require enterprise software. By combining free tools like Airtable with your existing automation platform, you can create robust review processes that ensure quality while saving time. This approach gives you the best of both worlds: AI efficiency with human oversight.

‍

If you’d like to learn more about automating your daily work with tools like Make and Airtable, be sure to explore the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

‍

In October 2024, the most innovative and exciting companies in the no-code and low-code space gathered in Paris for the annual NoCode Summit.

‍

Several members of XRay.Tech’s team attended to learn about the latest developments in no-code, share our services, and get to know the people in our industry. 

‍

At the conference, XRay’s Doc Wiliams had the chance to talk with Sean Montgomery, co-founder and CTO of Xano, a no-code backend database focused on building through visual development. 

‍

They discussed the evolution of visual development, Xano's journey, and what's on the horizon for the company.

‍

Here are some highlights from their conversation. 

‍

‍

From Google Photos to visual development pioneer

‍

Sean Montgomery's journey in technology spans 25 years, beginning with notable roots at Google. As one of the lead engineers on the original Picasa team and later Google Photos, Sean spent six and a half years at the tech giant before venturing out on his own.

‍

"After that amount of time, I got my confidence to leave the nest," Sean explains. He went on to form a development agency, which he ran for about ten years. It was during this period that Xano began to take shape, initially as an internal tool.

‍

"In the beginning, it was an internal tool just for us. Kind of like our secret weapon," Sean recalls. "But eventually it got so good year after year, with client-driven requests – we weren't making stuff up. The clients were demanding it, so it kept getting more efficient by client-driven demand."

‍

What started as a strictly internal tool became the basis for Xano, a platform that makes it easy to create a no-code backend for an application. 

‍

Evolution from no-code to visual development

‍

Today, Xano positions itself as the leader in backend visual development. Sean believes visual development represents the evolution of no-code, addressing some of the stigma associated with traditional no-code solutions.

‍

"No-code kind of has a stigma where it's not good enough for traditional developers, but as it keeps evolving, it's clearly working. The productivity gains are through the roof, and it's kind of hard to ignore," Sean explains. "So we believe it's time to rebrand the movement."

‍

‍

Sean and the team at Xano believe that the term “visual development” better encapsulates the sophistication and capability of modern no-code tools, and makes it clear that the technology doesn’t need to exclude traditional developers at all. 

‍

Transforming business through standardization

‍

One of Xano's key benefits is its ability to standardize development processes. Sean emphasizes how this standardization addresses a common challenge in traditional coding environments: "When you're talking about pure code, everyone kind of does it their own way, and everyone may be smart, but one developer does it their way and another one doesn't understand it."

‍

Visual development solves this by providing consistent, visual representations of logic flows. "The nice thing about visual development is it always looks the same. You can see the flow diagrams, the loops. It means that it lowers the barrier of entry for other people to contribute – non-engineers, project managers. It's really great."

‍

Real-world impact: a case study

‍

Sean shared a compelling example of Xano's impact through their work with Heimstaden, a large enterprise company in the rental apartments sector. The company faced challenges with scaling their development team while maintaining quality.

‍

‍

"They have a large development team, but they needed to do more. And the issue is…hiring good developers is tricky," Sean explains. "So they needed to actually take a step back and become more efficient with less. They went from all-code to a no-code flow where they're using Xano... now they're actually able to do way more with way less and are moving faster than they ever dreamed."

‍

Looking Ahead: Xano Script and AI Integration

‍

Looking toward 2025, Xano is introducing exciting new developments, particularly with Xano Script. This new feature bridges the gap between visual and traditional development, offering a new way to add code to a backend database. 

‍

"We're making [Xano Script] literally look exactly like the visual development," Sean explains. "The structure is the same, which means that we spent a lot of time trying to compact a lot of information, but without making it cryptic."

‍

The vision extends beyond just creating a new scripting language. Xano plans to leverage AI to transform Xano Script into functional syntax in other programming languages and vice versa. 

‍

“What we want to do is have the Al scan Xano Script and then actually turn that into other languages. So if you feel like ‘I can do amazing things with Xano, but I have this Python project where I need that’ – No problem.”

‍

This two-way conversion capability could be particularly valuable for companies dealing with legacy code.

‍

"There is so much legacy code that is so old but critically important for enterprise companies... Everyone's scared to touch it. Well, let's bring it into Xano and introduce that old code to visual development. All of a sudden, because the barrier of entry has been lowered, people can now understand what's going on and then start actually changing it instead of being scared of it."

‍

With their visual development tools, Xano aims to give every company the ability to build and update effective products. 

‍

---

Ready to transform your business with Xano or other no-code and low-code tools? XRay specializes in helping companies build efficient, scalable solutions using visual development platforms. Learn more about our services here on our site and schedule a free consultation call when you’re ready. 

‍

You can also explore our blog or our YouTube channel to learn more about automating your team’s daily work. 

‍

This interview was conducted by Doc Williams from XRay.Tech at the NoCode Summit 2024 in Paris.

‍

Are you tired of automation workflows that can only follow rigid, predefined paths? What if you could create an AI assistant that adapts to different situations and makes intelligent decisions on its own? Thanks to Zapier's new AI agents feature, you can now build sophisticated automated workflows that learn from your data and handle complex scenarios - all without writing a single line of code.

‍

Understanding Zapier AI Agents

‍

Traditional Zapier automations (called "Zaps") follow a largely linear trigger-action-action model, with optional branching logic. 

‍

While versatile and effective, these workflows can become complicated when you design them to handle multiple different scenarios. Each new case requires additional branches and conditions, leading to complex decision trees that can be difficult to maintain.

‍

‍

Zapier AI agents take a different approach. Instead of explicitly programming every possible path, you simply provide instructions in plain English along with relevant data sources and modular automated actions for the AI to run.

‍

‍

The AI agent then interprets incoming triggers (like emails), analyzes available data, and takes appropriate actions based on its understanding of your instructions.

‍

Creating Your First AI Agent

‍

Let's walk through building a basic customer support AI agent that can handle incoming email inquiries.

‍

Starting with a Template

‍

1. Sign in to your Zapier account and click on the “Agents” item in the left hand menu.

‍

‍

2. Select "New Agent".

‍

‍

3. Create an agent from scratch, or select a template from the list (e.g., “Support Email Agent”)

‍

4. Click “Configure” (if you’ve chosen to start from scratch)

‍

Configure your trigger

If you’ve chosen a template, you’ll immediately see your agent’s first “behavior” (or set of instructions to follow and actions to take). You can also access this if you’ve started from scratch by clicking on “configure”. 

‍

Like a normal Zap, every agent behavior must begin with a specific trigger. You can choose from any of the triggers available among Zapier’s 7,000+ integrations. 

‍

For the Support Email Agent, the template will default to using a “Gmail: New Email” trigger, but you can change this to anything you’d like (such as a different email provider like Outlook, or  a totally different app like HubSpot). 

‍

‍

Configuring the trigger for agent behavior works just like configuring a trigger for a normal Zap. Once you’ve chosen your app and trigger event, you just need to sign in to the corresponding app and configure your settings. 

‍

‍

However, you won’t need to test the trigger just yet. Instead, proceed to reviewing and editing the instructions.

Writing instructions for your agent to search for data and perform actions

‍

The instructions panel is where you define your agent's behavior. Unlike traditional Zaps, you write in natural language what you want the agent to do. For instance, in the Support Email Agent template (pictured below), the initial instructions read:

When I get a new email, look at the contents and determine if it is a question from a customer. If it’s a customer question, look up the answer in [use "Find data" to add a data source here]. If you find an answer to their question there, (Gmail: Create Draft Reply) to the customer in a direct, courteous tone with the correct answer. Please be succinct.

Important: Do not attempt to answer the question if you do not see a good answer in the attached data source. Do not make up information and do not try to be helpful when you do not know the correct answer. 

You can customize these instructions however you’d like. For instance, we’ll add some text at the end about sending a Slack notification for any drafted emails. 

Once you have created a draft, send a DM in Slack with a link to the draft. 

‍

Adding data sources

‍

In order to efficiently and accurately provide answers, summaries, or suggestions, your AI agent needs information to work with. Click on “insert data” to connect a new data source. 

‍

‍

You can connect directly to data in several popular apps like Notion, Airtable, Hubspot, and Zendesk. Alternatively, you can upload a file directly. 

‍

‍

Note that you can add multiple sources to a single agent behavior. 

‍

For our example, we'll upload a simple FAQ document that the agent can reference when answering customer questions.

‍

‍

Once you’ve uploaded your source, make sure to place it where you want within your prompt. 

‍

‍

You can drag and drop the data source pill to move it where you’d like, although we had a few issues with this functionality during our tests. If the pill isn’t moving when you click and drag, try cutting and pasting it instead. 

‍

Configuring Actions

‍

One of the best features of Zapier AI agents is their ability to perform automated actions in thousands of different apps, all while referencing the information in your prompt and data sources. 

‍

Common actions include:

• Drafting email replies

• Creating records in connected apps

• Sending notifications

• Adding labels or tags

‍

What’s especially useful is that the agent can dynamically generate values for any field in these actions based on its analysis.

‍

To add an action, click “insert action”.

‍

‍

Choose the app you want to use, such as Slack. 

‍

‍

Choose your desired action, and the action will be added to your instructions.

‍

‍

You can then click on the action to configure it however you’d like. 

‍

‍

You’ll need to pick an account, but the vast majority of fields can potentially be filled in by your agent. 

‍

However, if you want to configure a setting to a single static value (or to a static set of a few possible values), then we’d recommend making that selection to put appropriate guardrails on your AI agent. 

‍

‍

For our Slack example, we’ll set a static recipient for the DM, but the rest will be left up to the agent. This means that its responses may vary whenever it runs, but it’s also extremely easy to set up. 

‍

‍

Once you’ve finished setting up your action, click “Save”. 

‍

Note: you will also need to configure any actions that have already been inserted in your chosen template. 

Testing your Zapier Agent

After you’ve finished adjusting your instructions, save your work. Then, you can give your agent a try by clicking on “Test behavior” in the “Behavior Preview” panel. 

‍

‍

Your agent will then test your trigger and process your instructions with the data it retrieved. As it does its work, it will show every step of its progress and decision making. For more detail, just click on any module with a toggle.

‍

For instance, expanding the “Trigger” module shows the data the trigger retrieved. 

‍

‍

Once the agent has completed an action, it will show you a preview of the result and give you the option to approve it before it proceeds. 

‍

‍

‍

After you approve a preview, the Agent will actually run the action.

‍

‍

After a successful test, you can turn on your agent’s behavior to have it run in the background. It will launch whenever your trigger event occurs - no additional prompts or chats required.

‍

‍

Advanced example: building a multi-purpose support agent

‍

Let's wrap up this guide with a quick look at a more sophisticated example that handles three different types of incoming emails. 

‍

These instructions (pictured below) provide the agent with details, data, and actions to handle lead inquiries, questions from current customers, and junk email. 

‍

‍

This single agent effectively replaces what would have been a complex web of traditional automation branches, making it easier to maintain and modify as needed.

‍

Easily accessible AI agents

‍

Zapier's AI agents represent a significant step forward in making automation more accessible and powerful. By combining natural language instructions with AI decision-making capabilities, these agents can handle complex workflows with simple instructions, and adapt to variations in incoming data without needing to create intricate branching paths. 

‍

Whether you're managing customer support, processing leads, or handling internal workflows, Zapier AI agents offer a more flexible and intelligent approach to automation - no coding required.

If you’d like to learn more about optimizing your team’s daily work with AI and automation, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

If you want to use AI in your company, you might want to consider ditching the chatbots and connecting AI directly to your data instead. With Airtable AI, you can integrate AI-driven automation right into your database, making it more scalable, efficient, and tailored to your needs.

Airtable AI lets you write prompts with variables that pull data from specific fields in your records. Instead of manually entering information into an AI chatbot, you can automatically run AI-powered tasks for every record in your base—filling in the blanks with unique data from each row.

Unlike traditional chatbot interactions, this approach scales effortlessly. You can process thousands of records through a single, preformatted AI prompt, making it a powerful solution for automating workflows, generating content, summarizing data, and much more.

In this post, we’ll show you how to get started with Airtable AI fields step by step. 

Adding Airtable AI to your account

Airtable AI is available as an add-on for any Airtable plan. Users can try it out with 500 free credits for AI-generated content. 

‍

Once your trial credits are used up, you can add Airtable AI to your subscription for an additional $6/user/mo. 

‍

You can learn more about Airtable AI pricing here. 

Creating an Airtable AI field

To make an Airtable AI field, add a new field to any table, or edit an existing one. Select “AI” as the field type.

‍

‍

You can then choose the specific field type. There are currently four text-related options for AI fields in Airtable: Generate text (long text), Suggest records to link to (Linked record), Tag records (Multiple select), and Categorize records (single select). 

‍

‍

Note: After we recorded the companion video for this tutorial, Airtable also added a few AI options related to numbers. We’ll cover these in a future post. 

‍

After adding an AI field, you can configure a few technical settings and draft your prompt. We’ll focus first on a deep dive into the Generate text field; then, we’ll wrap up the guide with a brief overview of the other field types. 

Airtable AI text field technical settings

You can access an Airtable AI field’s technical settings by clicking on the gear icon on the right of the field details panel. 

‍

‍

Generate automatically

When enabled, the “generate automatically” setting will make your AI field regenerate its response whenever a record is updated with new data. 

‍

‍

Use caution with this setting. Enabling it could cause you to quickly burn through your available credits if your databases are updated frequently. In most cases, we’d recommend leaving it disabled. 

‍

Model

Airtable AI fields let you choose from a variety of models made by different AI organizations. Each model has its own strengths and weaknesses, and will use credits at different rates. For easy reference, certain models are labeled as “low-cost”, so you can pick these if you’d like to optimize for price efficiency. 

‍

‍

You can select models created by OpenAI, Anthropic, Meta, and even IBM. To learn more about each model, you can refer to the documentation created by each developer. 

‍

Randomness

The “Randomness” setting determines how random or consistent the AI’s output will be, with options for Low, Medium, and High. Higher randomness will generally result in more varied and creative output, while the default choice of “Low” will keep the AI’s responses consistent and more reliable. 

‍

‍

Composing your Airtable AI prompt

After you’ve configured your technical settings, you can draft the prompt for your AI field by either choosing a template or starting from scratch. 

‍

‍

In the “AI prompt” field, you can enter text instructions directly just like you would with a chatbot like ChatGPT or Claude. For instance, in our example below, we’ve started a simple prompt about composing a blog post. 

‍

‍

Note that you can also reference other fields as variables within the prompt. 

Referencing other fields

Type { or click on the plus sign to instruct the AI to access the fields in your table. 

‍

‍

When the AI prompt runs for each record, it will refer to the values in your chosen fields within the same record. So in our example below that references the “Content Title” field, Record 1’s prompt will refer to Record 1’s Content Title. Record 2’s prompt will refer to Record 2’s Content Title, and so on. 

‍

‍

This dynamic data is what makes Airtable AI much more scalable than a simple standalone chatbot. 

‍

Previewing the AI Output

The preview window beneath your prompt will show you a sample of how the AI will respond to your current prompt. It will also tell you how many credits each run of your prompt will consume, with longer prompts and responses consuming more credits. 

‍

Generating your prompt with AI

If you’re not sure exactly how to craft your request for Airtable AI, you can try generating the prompt itself with AI as well by clicking on “Generate Prompt”. 

‍

‍

Just enter your initial idea for the prompt in the window, and press “Generate”. 

‍

‍

Then, Airtable AI will expand on your prompt with helpful detail and added structure. 

‍

‍

You can then tweak the AI-generated prompt however you’d like. For instance, we’ll add another field (“Content Description”) to the “Context and Data” section of the prompt. 

‍

‍

Once you’ve finished adjusting your prompt to your liking, click on “Save” to commit your choices and close the field. 

Generating responses with your Airtable AI fields

If you’ve chosen to generate answers automatically, then responses will generate for every record in your table. As noted earlier, this automatic generation can quickly consume large amounts of credits, so be careful and only enable it when necessary. 

‍

If automatic generation is disabled, you can generate a response for your chosen record by clicking on the “Generate” button in its AI field. 

‍

‍

After a moment, the AI will enter its response into the field. 

‍

‍

While clicking manually to generate a response is certainly a bit more work than fully automatic generation, it’s still far more efficient than copying and pasting the same prompt dozens of times for each record. 

‍

If you’d like to edit your AI prompt settings, just edit the field at any time. 

Other AI field types: suggest linked records, tags, and categories

Now that we’ve covered text-generating AI fields in depth, let’s take a quick look at the other three options available: suggesting linked records, providing multiple tags, and choosing single categories.

‍

Suggest linked records with AI

With the “Suggest a record to link to” option, you can get the AI’s recommendation for a relevant record from a linked table. 

‍

For example, our “Content” table is linked to a “Themes” table, where each record describes a common theme in our content. We’ll use this AI field to recommend appropriate themes for each piece of content. 

‍

‍

First, you’ll need to choose which table the AI should search in. We’ll choose the “Themes” table. 

‍

‍

Then, you’ll see the usual options for configuring a linked record field. You can choose whether or not to allow multiple records, limit the selection to a specific view, or filter the available records by a custom condition.

‍

‍

The last choice about “Top matches” is just what makes this an AI-powered field. If you disable it, the field won’t pick linked records with AI - it will just be a normal, manual linked record field. 

‍

Finally, you can pick the fields that the AI should search in both tables when finding records that match. In our example, we’ll have it compare the Content Title and Content Description to the Theme Name. 

‍

‍

Click save to commit your choices. Then, to get the AI’s recommendation for any given record, just click on the plus sign.

‍

‍

At the top of the list of records that appears, you’ll see the choices that the AI considers to be the best fits. You’ll still have to select them manually, but the AI should do a good job of picking the most likely options out of your entire table. 

‍

‍

Tags and categories

The last two choices for text-related AI fields are essentially the same, but one relates to multiple select fields while the other handles single select fields. 

‍

‍

With AI-chosen tags and categories, you can get the AI’s help in selecting an appropriate option from your preconfigured single and multiple select fields. They’re configured in largely the same way, so we’ll just cover the “Tags” option. 

‍

First, you’ll need to choose which specific fields the AI should reference to make its evaluation. In our example, we’ll just have it look at the ‘Long-form Blog’ field (which is the full blog post we drafted with AI earlier). 

‍

‍

If you don’t already have options for this multiple select field, you’ll need to add them here. The AI will just choose from your preconfigured options - it won’t create new ones. 

‍

We’ll add a few tags for topics and formats that come up a lot in our blogs: Tutorial, No-Code, Low-Code, and AI. 

‍

‍

Once you’ve created or updated the field, you can click “Generate”, and the AI will fill in one or more tags for your chosen record. 

‍

‍

‍

The “Category” option works in exactly the same way, but for single select fields. 

‍

Choosing the right AI tool for your business

Tools like ChatGPT are perfect for ad hoc, one-off AI tasks. However, integrating AI into your company’s workflows requires a more systematic approach that lines up with your structured data. 

Fortunately, Airtable AI makes it easy to plug your company’s data directly into customized AI prompts. Give it a shot today to see what it can do. 

If you’re looking for a partner to support your company in its AI and automation efforts, reach out to XRay and book a free 15-minute consultation to discuss your options.

If you’re getting tired of running the same prompts in ChatGPT day after day, then ChatGPT’s latest beta feature is going to make your life much easier. 

‍

With the newly released “Scheduled tasks” feature, ChatGPT users can schedule their favorite prompts to run automatically. Whether you want daily news summaries, language practice, or custom AI workflows, you can now set up the prompt once and schedule to repeat it as often as you want. 

‍

In this guide, we’ll show you how to use this game-changing feature and share our opinions on what still needs work. 

Where to find Scheduled Tasks in ChatGPT

While scheduled tasks remain in beta, you’ll need to have a Plus, Pro, or Team plan with OpenAI to access them. 

However, OpenAI plans to roll out the feature with more users over time, and it’s likely to be available to all users before too long. 

Starting a new Tasks chat

To create a new Task in ChatGPT, open up a new chat. Then, go to the model selection menu in the top left of the window. 

‍

Choose “ChatGPT-4o with Scheduled Tasks”.

‍

‍

Note: sending prompts to this model will count against your plan’s 4o usage limits. 

‍

You can also access scheduled tasks by clicking on your profile icon, and selecting “Tasks”. 

‍

Once you’ve changed the model, you may see a few options for creating preformatted tasks. You can also write a prompt to create your own custom task. 

‍

‍

Creating a custom Scheduled Task in ChatGPT

To create a custom scheduled task, just write a prompt explaining what you want ChatGPT to do, and how often you’d like it to repeat the task. 

‍

For example, you might give ChatGPT a task like this:

‍

Every day at 10:05 AM, search the web for the most interesting no-code news and give me a summary of what you find. Include links to each story or post. 

‍

‍

After you send your prompt, ChatGPT will reply with a confirmation of your newly scheduled task. 

‍

‍

You can click on the three dots menu to edit the details of your task. You can change the task title, the prompt text, and the scheduling. 

Where to find ChatGPT scheduled task results

ChatGPT won’t respond to your task prompt until the scheduled time, but you don’t need to leave your browser open on ChatGPT while you’re waiting. 

‍

By default, ChatGPT will send you push notifications and email alerts whenever the chatbot has completed one of your scheduled tasks. 

‍

‍

You can enable or turn off either form of notifications by clicking on your account profile image in the top right, selecting ‘Settings’, and choosing ‘Notifications’. 

‍

‍

Refining your task

After ChatGPT completes your first scheduled task, be sure to review the results carefully. If there’s anything you’d like to change, you can always update the prompt to better guide the chatbot in the direction you’d prefer. 

‍

For example, you may need to specify a particular format that you’d like to use, or information that ChatGPT should always include. In our earlier tests, ChatGPT didn’t consistently include links in a task to summarize daily tech news headlines. 

‍

‍

‍

The task that ChatGPT saves will be a summary of your original request, and may not include all of your desired details, but updating the prompt in the task should help you get the results you’re looking for. 

Limitations and issues with ChatGPT tasks

As we noted earlier, ChatGPT tasks are still in beta. As such, you should expect to encounter some technical issues while you’re setting up tasks for a while. 

Sometimes slow to respond

For instance, the very first task we configured simply didn’t run at its scheduled time. After rescheduling the task a couple of times, ChatGPT finally responded on time. 

‍

‍

Inaccurate or unclear information

In another test, we set up a task for ChatGPT to give us a summary of Denver, CO’s weather each morning. 

‍

While ChatGPT produced a beautifully formatted meteorological report, the actual data seemed a bit suspect. 

‍

‍

It reported a temperature of 14°F at 9:00am MST on January 21st 2025, but clicking on its source link showed a temperature of 28°F at the same time. Google also showed a temperature of 24°F.

‍

‍

Errors like this might be related to a software bug, might they might also just be an issue with generative AI in general. LLMs aren’t ideal for remembering facts and repeating them precisely. Instead, they have a tendency to transform, embellish, or otherwise alter information. 

‍

If you don’t want any creativity entering into the workflow (like summarizing, interpreting, etc.), then it may be better to move data around with a no-code automation instead. 

‍

Building no-code automations with XRay

If you’d like to build an AI and no-code workflow tailored to your team, just reach out to XRay for a free consultation. 

‍

We’ll spend 15 minutes addressing any no-code or low-code problem you’re having, and then we’ll go over your options for a full project or membership.

‍

Our clients and members save hours on their recurring tasks, reduce human error, and have an easier time finding all the data they need. 

‍

Just click here to schedule a call. 

Streamlining your AI workflows with scheduled tasks

Scheduled tasks in ChatGPT are a promising feature, and with some good updates to make them a bit more stable, this is going to become a great way to schedule your recurring prompts. 

If you don’t have access to the beta yet, you might want to check out our tutorials for automating your prompts with automation providers like Zapier and Make. Also be sure to check our blog each week for new tips, tutorials, and updates about no-code tech and AI.

Like many SaaS platforms today, Slack has developed its own native automation capabilities. Available on Pro plans and above, Slack Workflows allow users to automate routine tasks directly within their workspace. 

But how do these built-in features stack up against dedicated automation platforms? Let's dive into what Slack Workflows can do, and explore their strengths and limitations.

The Strengths of Slack Workflows

1. User-Friendly Interface

One of Slack Workflows' standout features for no-code builders is its approachable interface. The platform uses simple, plain language throughout the automation builder, making it accessible to users with minimal technical background. Instead of technical terms like "triggers" and "actions," you'll find friendly phrases like "start the workflow" and "then, do these things."

‍

2. Seamless Integration with Slack Features

As you might expect, Slack Workflows excel at automating tasks within Slack itself. The platform offers robust options for triggering workflows based on people joining channels, message reactions, and other Slack-specific events.

‍

3. Webhook Support and External Connectivity

Despite being a built-in tool, Slack Workflows offers surprising flexibility in connecting with external systems. The platform includes webhook triggers, allowing you to initiate workflows from external applications. 

‍

This feature significantly expands the potential use cases for Slack automations, giving you the option of triggering your automations with the click of a button from tools like Airtable. 

4. Airtable search integration

One particularly valuable feature is the surprisingly useful Airtable integration, which includes basic search functionality. Your choices for automating Airtable in Slack aren’t as extensive as they are in providers like Zapier and Make, but the ability to find Airtable records still adds considerable depth to your Slack automations. 

‍

‍

This allows you to use Airtable as an operational database to feed data into your automations and create dynamic workflows. 

‍

‍

For example, you can use Airtable lookups to systematically route automatically created tasks to different projects or send messages to specific users based on custom criteria.

‍

‍

5. Built-in Forms

Slack Workflows includes a convenient form builder that allows you to collect data from users without redirecting them to external URLs. These forms integrate seamlessly with the Slack interface and can feed directly into your automation steps.

‍

Areas for Improvement: Where Slack Workflows Fall Short

1. No Testing Capabilities

One significant drawback is the lack of testing features. Users cannot test individual triggers or steps within a workflow, making it necessary to run the entire automation to verify that it works as intended. This can make the development process more time-consuming and frustrating, especially for complex workflows.

‍

By contrast, automation providers like Zapier, Make and Pipedream all include easy built-in testing. 

2. Limited Integration Options

While Slack Workflows does offer some external connectivity, its integration library is relatively small. With less than a hundred available app connections, it falls far short of the thousands of integrations offered by dedicated automation platforms like Zapier, Make, or Pipedream.

‍

3. User Experience Issues

Based on our use, several minor but noteworthy UX issues exist within the platform. For example, while the confirmation buttons for launching automations are helpful, they lack basic options like "Cancel" or "Stop." These small oversights can add friction to the user experience.

‍

4. Limited Scalability

As with many built-in automation tools, Slack Workflows faces challenges in terms of scalability and maintainability compared to standalone automation platforms. The limited testing features and integration options can make it difficult to build and maintain complex automated systems that connect several different tools and the teams using them.

Our Verdict

Slack Workflows can be a valuable tool for teams looking to automate simple, Slack-centric tasks. The platform shines when it comes to use cases like:

• Creating quick automations for task management
• Scheduling meetings
• Composing emails directly from Slack
• Building simple workflows that leverage Slack's native features

However, for organizations seeking to build comprehensive automated infrastructure or connect multiple applications to a centralized source of truth, standalone automation platforms remain the better choice. 

Tools like Zapier, Make, and Pipedream offer greater scalability, more robust testing features, and significantly more integration options and are a much better fit for enterprise organizations seeking to automate their workflows. 

If you’d like to learn more about automating your team’s work with a variety of no-code and low-code tools, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Imagine describing your dream business application in plain English and watching it come to life in seconds. That's the magic of Airtable's new AI Cobuilder, a breakthrough feature that transforms natural language prompts into fully-functional custom applications. 

Whether you need a sophisticated content management system, a project tracker, or an inventory database, AI Cobuilder can create it faster than you can pour a cup of coffee.

The possibilities are endless: describe what you want to track, how you want to view it, and what relationships matter to your business. AI Cobuilder handles the technical heavy lifting, creating tables, establishing relationships, and even generating intuitive interfaces.

Let's explore how this powerful tool works using a content management system as our example.

Understanding AI Cobuilder's Capabilities

‍

To get started with Airtable’s AI cobuilder, just open up Airtable and create a new base. You’ll see a choice to either build from scratch, or build with AI. 

‍

‍

From there, all you need to do is enter a prompt to ask for what you want. To get started quickly, you can click on any of Airtable’s app suggestions like a Content Calendar, a Task Management portal, or an Analytics Dashboard. 

‍

However, you can also write a prompt to ask the Cobuilder to create whatever you’d like. The AI Cobuilder opens up endless possibilities for custom applications. The key is knowing how to guide the AI to create exactly what you need.

‍

Crafting the Perfect Prompt

‍

The secret to getting the most from AI Cobuilder lies in how you structure your prompt. Here are a few key tips you can apply to your prompt for any application you want to build

‍

1. Define Core Purpose: Start with a clear statement of what you're building. For example: "Create a base dedicated to managing our company's content pipeline."

‍

You may also want to include details about your company, or the team that’s going to be using the base. 

‍

2. Specify Data Requirements: List the key types of information you need to track. In our content management example, this included platforms like LinkedIn and YouTube, along with file uploads for content.

‍

This data will help Airtable to populate your select fields with relevant options. 

‍

3. Establish Relationships: Describe how different pieces of data relate to each other. Our example linked content pieces to themes and campaigns.

‍

The cobuilder will be able to create dynamic links between your tables based on your request, so be sure to include this information in your prompt. 

‍

4. Detail Specific Fields: Be explicit about what fields you need • from simple text fields to more complex requirements like file uploads or multi-select options.

‍

Naming the field type will ensure that Airtable builds exactly what you’re looking for. 

‍

5. Request Desired Views: Specify how you want to visualize your data, such as calendars, kanban boards, or custom dashboards.

‍

These views will be constructed in your base’s Interface, which is typically where you and your team will spend most of your time interacting with your new app. 

Our prompt: building a content management base

Below, you can find the prompt that we used to create a content management base in this video. You may want to consider copying and pasting it as a starting point. 

‍

Create a base dedicated to managing our company’s content pipeline. 

The base should maintain a centralized repository for all content posted on multiple platforms; LinkedIn, Facebook, X, Tiktok, and YouTube. 

‍

Each piece of content should link to a Theme in a separate table. For example some themes include Automation, Workflow Design, Tutorial, and Strategy. 

‍

Each content campaign should be associated with a long-form blog, a video (file upload), raw footage (file upload), title, description, meta tags, and SEO tags. 

‍

Include 3 AI fields for content  campaigns: 1 to draft a post on the social media of choice, 1 to generate SEO tags, and 1 to identify who the best audience persona is for the content.

‍

Include a calendar to easily schedule upcoming content. 

‍

In the “Themes” table, include a count of the number of pieces of content, along with a lookup of all the personas that are targeted.

‍

NOTE: The Airtable Cobuilder won’t create AI fields for you, even if you request them specifically like in this prompt. However, it will create placeholder text fields. 

Adjusting the Cobuilder’s Output

After you enter your prompt, the cobuilder will take a moment to build a preview of your new app. The preview will only include an interface – you won’t see the core data layer here. 

‍

‍

However, you can interact with the preview as though it were a finished app. Look it over and see if there’s anything you’d like to adjust. 

‍

Just click on Airtable’s suggestions to make a few quick additions, or select “Add your own” to enter a custom request. 

‍

In our example below, selecting “Assets” and “Personas” adds pages for both when the preview is regenerated. 

‍

‍

When you’re happy with the preview, click on “Create this app” to produce an actual Airtable base in your workspace. 

‍

Note that you can also edit the base manually once it’s been created. 

‍

What AI Cobuilder Delivers

‍

Here’s what you can expect from an AI-built base in Airtable. 

‍

• Comprehensive Data Structure: Tables and relationships are automatically created based on your requirements

‍

• Sample Data: The Cobuilder generates a few records of mock data to help you understand how the system works

‍

‍

• Intuitive Interfaces: Custom views and dashboards for different aspects of your application

‍

‍

• Smart Additions: The AI often includes helpful features you might not have explicitly requested. For instance, the Cobuilder provided a couple of forms that we didn’t request in our example prompt. 

‍

Current Limitations:

• No automation setup

• Some specialized fields (like AI fields) need manual configuration and cannot be created by AI

• Basic interface layouts that may need refinement

‍

‍

Beyond Content Management

‍

While we've used a content management system as our example, the same principles apply to building:

‍

• Project management tools

• Customer relationship databases

• Inventory tracking systems

• Event planning applications

• Resource scheduling tools

• And much more

‍

Making the Most of Your Generated App

‍

After generation, take time to:

• Adjust interface layouts to match your workflow. To begin editing your interface, just press Option 1 on Mac, or Alt 1 on Windows. You can learn more about interfaces here in our Airtable beginner’s guide.

‍

• Add any missing fields or relationships. The Cobuilder will typically fulfill most of the specific instructions in your prompt, but may still overlook a few details. 

‍

• Set up necessary automations. The Cobuilder will not build any automations for you, so you’ll need to set them up yourself. You can explore several tutorials on our blog to learn more about automating your work with Airtable. 

‍

• Customize views and filters. Airtable offers a wide array of customizable views and filters for your interfaces and data. While the Cobuilder will likely provide a lot of helpful views in the interface, you’ll probably need to handle all of this yourself in the data tab. 

‍

• Configure user permissions. With Airtable’s granular user permission controls, you can easily determine which users can view, edit, and add data. 

The Future of App Building

‍

AI Cobuilder represents a fundamental shift in how we create custom applications in tools like Airtable. What once took hours or a week of work can now be accomplished in minutes, allowing teams to focus on refining and optimizing rather than basic setup.

‍

While it's not a complete replacement for expert setup, AI Cobuilder excels at rapidly prototyping ideas and creating functional first versions of Airtable applications. As the technology evolves, we can expect to see even more sophisticated features and capabilities added to this already impressive tool.

‍

If you’d like to learn more about streamlining and automating your work with no-code tools like Airtable, be sure to explore check out our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

‍

Forms are one of the most common starting points for automation. They’re an easily accessible and shareable way to collect the data that drives our automated workflows. 

‍

In this post, we’ll explore Jotform's newly updated Workflows feature, which combines robust form building with native automation capabilities.

‍

Note: This post and its associated video are sponsored by Jotform. However, all opinions expressed are our own.

‍

What is Jotform?

‍

Jotform is much more than your standard form builder. While it includes all the basic fields you'd expect, it also offers advanced features like payment processing, digital signatures, and CAPTCHA integration. 

‍

‍

Using its drag-and-drop interface, you can create everything from simple contact forms to full-featured storefronts.

‍

‍

Building an automated expense reimbursement workflow

‍

Let's look at how to create a practical automated workflow in Jotform through a real example: an expense reimbursement process.

‍

Our workflow will:

1. Collect expense details and receipts through a form

2. Route the request to a supervisor for approval and signature

3. Upload approved requests to Google Drive

4. Create follow-up tasks for software-related expenses

‍

‍

Here's what makes this particularly powerful: 

‍

Jotform can handle the entire approval process in a single workflow. Unlike other automation platforms where you might need multiple workflows for a multi-step approval process, Jotform lets you build the complete flow in one place.

‍

Key features

Approval management

• Built-in approval routing

• Digital signature collection

• Ability to request additional information

• No time limits on approval waiting periods

Conditional logic

You can create different paths based on form responses. In our example, software-related expenses trigger additional tasks automatically.

Integrations

While not as extensive as dedicated automation platforms, Jotform connects with popular services like Dropbox, Salesforce, HubSpot, Mailchimp and Google Drive – which we’ll use in this automation. 

‍

Step 1: Create your form

First, create a form with these essential fields. You can include additional fields if you’d like, but the ones listed here will be necessary for the automation as we’ve built it. 

‍

• Name

• Email address

• Supervisor's email

• Expense category

• Total cost

• File upload for receipts

‍

‍

‍

Step 2: Create a new workflow

Navigate to the Workflow Builder (click the arrow next to "Form Builder")

‍

‍

This will create a new workflow with just one element: “Starting Point: Form”. That means that your automation will activate whenever someone submits your form.

‍

‍

Step 3: Add approval process

With the “Approve & Sign” element, you can easily integrate an authorization step into your form workflow. 

‍

1. From the “Workflow Elements” panel, drag the "Approve & Sign" element into your workflow

2. Configure the settings:

   • Set "Approve" and "Deny" as the two possible outcomes

   • Enable signature requirement for the “Approve” outcome only

   • Set the reviewer as the supervisor (Select “form fields” and insert the supervisor email field)

   • Enable "Request More Information" in advanced settings

• Allows the supervisor to request additional info before making a decision 

   • Set completion rule to "Require response from one person"

‍

‍

The completed action will send a notification to the designated supervisor, asking them to approve or reject the request for reimbursement. 

‍

If the supervisor approves the request, they’ll need to provide their signature as well. 

Step 4: Configure the “Deny” path

Next, you’ll need to create automated actions for both the “Approve” and “Deny” outcomes that result from the approval step. 

‍

We’ll begin with the “Deny” path in our example, since it will be shorter – just one step alerting the original form submitter that their request has been denied. 

‍

‍

1. Click and drag to draw a new path out of the approval step 

2. Set the path to trigger for the "Deny" outcome (click “Select outcome”)

3. Add an "Email" action to the “Deny” path

4. Configure the email:

   • Set recipient to the original submitter’s email address

   • Customize the subject line with expense details (Click “Form fields” to insert data)

   • Use the template to include form submission details

   • Add a custom message explaining the denial

‍

Step 5: Set up the approval path

With the “Deny” path all set, you can now create an “Approve” path that will include automated actions that run when a request is approved. 

‍

‍

1. Draw a new path leading from the “Approve & Sign” step. 

2. Set the path to run for the “Approve” outcome (click “Select outcome”)

3. Add a Google Drive - Send Files step from the “Integrations” tab

4. Configure the Google Drive integration:

   • Connect your Google Drive account

   • Designate a folder to store form responses and uploaded files

   • Enable subfolder creation for each submission

   • Choose to upload both the submission PDF and receipt files

‍

‍

With these settings, the Google Drive step will upload a PDF summary of every approved submission, along with copies of any files that the user uploaded (e.g., receipts). 

‍

Click “Save” to complete the step configuration. 

Step 6: Add conditional logic

The automation could stop here, but we’ll add a few more steps to demonstrate Jotform’s conditional logic capabilities. 

‍

After a reimbursement request is submitted, the automation will check to see if the expense category is set to ‘Software’. If it is, then Jotform will create a task for the original submitter to add relevant credentials to the team’s password manager.

‍

If the category is set to anything else, then this task won’t be created. 

‍

‍

1. Draw a new path leading from the Google Drive step 

2. Add an "If/else condition" to the new path

3. Set condition: If "Expense Category" Is Equal To "Software"

Step 7: Add an automated action to create a task

Just as you needed to add “Approve” and “Deny” paths following an approval step, you’ll also need to add “True” and “False” paths following the conditional logic step. 

‍

‍

   • Draw a new path leading from the If/Else condition element.

   • Set the “Branch” to “True” (click “Select Branch”)

   • Add a "Task" element

   • Configure the task:

      • Set a clear title (e.g., "Add Software Credentials to Password Manager")

      • Write a detailed description explaining what needs to be done

      • Leave the outcome as default "Complete"

      • Set the assignee to the original submitter using form fields

      • (Optional) Enable email notifications and customize email

‍

Step 8: Complete the “True” branch by adding an email notification

Because your Task step has an outcome (i.e., the task is marked Complete), you need to provide a path that follows from that outcome. 

‍

In our example, we’ll just add a simple path that sends an email alert when the task is completed. 

‍

‍

1. Draw a path from the “Task” step

2. Set the outcome to “Complete” (click “Select Outcome”)

3. Add an "Email" action

4. Configure the email with any desired details

‍

‍

Since there’s no outcome associated with “False”, you don’t need to create a path in the workflow for “False”. The workflow will simply stop at the conditional logic step if the condition is not evaluated to be True. 

Testing your workflow

‍

Before using any workflow with your team, be sure to test it thoroughly by running through every step with test data. 

‍

For instance, to test this workflow, you would need to:

‍

1. Submit a test form response

2. Check that approval emails are received

3. Verify Google Drive uploads work correctly

4. Test both approval and denial paths

5. Confirm conditional tasks are created appropriately

6. Test with different expense categories

‍

Remember to test all possible paths through your workflow to ensure everything works as intended before sharing the associated form publicly or with your team. 

‍

As we tested our example workflow, we received an automatic request for a signature:

‍

‍

And the approved request was uploaded to Google Drive as a PDF:

‍

‍

Pricing overview

‍

Next, let’s review Jotform’s pricing, and discuss how Jotform can fit into your automation stack. 

‍

First and foremost, you should note that automated Workflows are included on all Jotform plans – even the free plan. 

‍

We’ll give you a brief overview of some of the key features available at each tier, but you can explore Jotform’s pricing and features in more detail on their pricing page. 

‍

‍

Free plan

The free plan gives you a lot of room to try out Jotform and explore its features. For instance, you can build everything we created in this tutorial using a free plan. 

‍

The main limitations on the free plan are general usage restrictions: you’re limited to 5 forms, 100 submissions a month, 100 MB of storage, and so on. 

‍

Each paid tier will expand on these limits, and will unlock a few unique features as well. 

‍

Bronze tier

For example, on the Bronze tier, starting at $39/mo, you can remove Jotform branding from your forms. 

‍

Gold tier

On Gold, which starts at $129 a month, you’ll be able to access HIPAA compliance features.

‍

Enterprise

On the Enterprise plan, which will require a custom quote from Jotform’s sales team, you can add multiple users to a single Team account. 

‍

Note that you can get a discount on all of these prices if you pay annually instead of monthly. 

‍

How Jotform Workflows fit into your software stack

‍

Jotform Workflows shines when you need to:

‍

• Build form-based automation quickly

• Handle complex approval processes

• Combine form features like payments and signatures with automation

‍

However, it's not meant to replace full-featured automation platforms like Zapier or Make. Instead, it works best either for form-centric workflows or in conjunction with other automation tools.

‍

Both Zapier and Make support Jotform, so you can build third-party automations in addition to the automations you build with Jotform workflows. 

‍

The ability to handle multi-step approvals in a single workflow is particularly impressive, and the native integration with Jotform's robust form features makes it a compelling choice for many business processes.

‍

What sets Jotform apart is how it combines advanced form features with automation capabilities. While it may not replace your existing automation tools entirely, it offers a unique and valuable approach to form-based workflows that could streamline many of your business processes.

‍

To get started with Jotform, you can create an account for free at jotform.com. 

‍

—

To learn more about automating your team’s daily work, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

2024 has been an exciting year for our company, and one of the biggest highlights for us was bringing amazing new talent on board. 

If you're a no-code builder on YouTube, you probably already know the name Doc Williams from his channel, where he posts excellent no-code tutorials, reviews, and more. But now he's added another title to his already impressive resume.

‍

Recently, Doc joined the team at XRay to become the program manager of LowCodeEngineers, XRay's service for matching people with expert low-code support on a convenient hourly basis. In October, Doc and XRay CEO Tom Nassr had the chance to chat for a bit at the No-Code Summit in Paris.

‍

They talked about Doc joining the team, how LowCodeEngineers changes the game for freelancers, and what they're looking forward to in the future. Here's the full conversation.

‍

‍

Interview: Tom Nassr and Doc Williams at NoCodeSummit 2024

‍

Tom Nassr

You're sporting the XRay.Tech flag now! How does that feel?

‍

Doc Williams

It feels good. I was a little bit nervous when the announcement happened, but I'm excited to be officially on the team. And so excited about all the things we're going to be doing with LowCodeEngineers.

‍

Tom Nassr

LowCodeEngineers is going to have a lot of work to do after this. Obviously, there's just people asking to join as a low-code engineer. There's also different companies coming up and saying, “hey, I could hire somebody for an hour or two.” 

‍

What are some things that you're excited about for LowCodeEngineers now that you're shaping the program?

‍

Doc Williams

A lot of things. Number one, I think helping the customers – helping all of these different businesses.

‍

But also, I think I'm more – or equally – excited with the engineers themselves. 

‍

I've been a freelancer. I've run an agency. I know trying to have both hats on: trying to build and build my business. That's very stressful.

‍

Tom Nassr

Very.

‍

Doc Williams

And so with LowCodeEngineers having all of the communication with the customer figured out, having all the tools with Guardian, having Debrief, all these different tools that were announced as well. That really helps the freelancer be at their best with their customer, focus on building and then leave everything else out of it. Done. Automated.

‍

So that's what I'm really happy about, because now I can just help find these great, amazing engineers that want that hourly rate locked in. They can just build, and then they can leave and do whatever they want with their life outside of it. 

‍

And they don't have to manage an agency and do all of these other things that are really stressful and time consuming.

‍

Tom Nassr

Well, it's totally different from Fiverr or Upwork or any of these other marketplaces. I mean, those marketplaces really create a competitive environment for engineers or talent generally. It's a race to the bottom. 

Clients often don't know exactly what all of these fancy tools can do. They don't know exactly how it can transform their business or what they even need to ask for.

‍

But LCE is a total shift in how work goes. Instead of the general market saying, “Hey, fend for yourself. You find the right person to hire. You vet them in a technology you don't understand, and you manage the project, LCE is saying “This person can actually help you.”

‍

That little bit of high-touch helps a lot, and LowCodeEngineers takes on the expense of actually doing the vetting and pairing. It’s a total paradigm shift for the client and the engineer.

‍

Doc Williams

I love how you brought that up because a lot of times people think, “Oh, I know what I want, I'm going to go to Fiverr or Upwork.” 

‍

But they don't know what question to ask to find out if the freelancer they’re evaluating is actually an expert. Yes, they can be in Fiverr Pro, or you can see that they have a certain number of reviews, which is excellent, but that just means that they're hitting the bare minimum. 

‍

A lot of times, the clients don't know how to articulate what they need for their project.

‍

And then they're mad at the developer or the engineer because it's like, “You weren't listening”. And the engineer is like, “Well, you asked me to do this, and I did exactly that.” But they weren't given the right context and the scope and all of those things.

‍

With LowCodeEngineers, before they even get that customer, the engineer gets all of the context. They get everything that they need to know to be successful to go into that session.

‍

Tom Nassr

A lot of that happens in the sales process, right? I'm wearing that hat right now. Hopefully I'm not wearing that hat forever, but I'm happy to hand that over to the next person. But the big difference is that the sales process on XRay's side is that we meet with everybody for 15 minutes.

If someone thinks they have a problem worth solving, we'll get on the phone with them for 15 minutes. When they explain their problem, they often frame it in terms of specific tasks - like 'I need to move X data into Y application.' But we help reframe the conversation to focus on their desired outcome: 'What output do you need? What are you trying to decide on or act on based on that output?’

And you know, there's usually a couple different ways to create that dish. We tell them, 'Hey, we have experts on the bench here who will give you the recipe and bake the cake for you. Let's focus on the style of cake you want, rather than making you direct it every single step of the way.’

You know, you'd never go to a restaurant and tell the chef how to cook your dish, right? It just wouldn't happen. And no-code/low-code is an emerging discipline that I think will follow suit.

‍

Doc Williams

Absolutely. Absolutely. No, this is just the beginning, and it really does feel great. I'm really excited to see what's going to be happening in the next couple of months or a year down the line. And to see what we can bring to market to really shift the industry with this product, with this service. 

‍

Tom Nassr

I think it's just going to be one client, one engineer at a time. It's going to be an exciting ride. 2025 is going to be fun.

‍

Doc Williams

Yeah. Good times.

---

Note: some portions of the interview have been edited or paraphrased for clarity.

---

If you’d like to learn more or book time with LCE, just check out lowcodeengineers.com to get started. 

To learn more about automating your work with no-code and low-code tools, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Have you ever wished your Notion databases could do more than just store information? While text fields and numbers are great for collecting data, your database can actually do much more – in fact, it can think for itself. 

‍

That's where Notion formulas come in. They're like giving your database a brain, automatically crunching numbers, combining text, and making smart decisions based on your data.

‍

In this guide, we’re going to show you everything you need to know to get started with Notion formulas. 

‍

You can follow along with our examples by copying this template database into your own Notion workspace. 

How Formulas Work in Notion

‍

To add a formula property to your database:

‍

1. Open your database and add a new property

2. Choose "formula" as the type

3. Give the property a name

4. Click "edit formula" to begin writing your custom formula

‍

‍

Your formula can be just about anything you want it to be, as long as it contains at least one valid expression. 

An expression will typically consist of one or more functions, but could be as simple as a string of text that says "Hello world!". 

‍

‍

For example, if you create a formula with just the text "Hello world!", you'll see that exact text in this property for every page in the database. 

‍

‍

While neat, this “formula” isn't particularly useful without any functions. 

‍

Let’s replace it with a simple function: id()

id()

‍

‍

The id() function will fetch the unique page ID associated with every record in the database. 

‍

‍

Exposing this ID can be extremely helpful when you’re automating your Notion database with tools like Make and need to look up a specific page. 

‍

Let’s take a look at some more advanced functions that will let your formulas automatically edit text, run mathematical calculations, format dates, and process conditional logic.

‍

Working with text in Notion formulas

‍

Let's start with some functions that revolve around editing text strings. One common need is extracting a subset of text from a longer string. Here's how to create a formula that grabs the first three characters of a product name to start creating a product ID:

‍

1. Add a new formula property and edit the formula

2. Type "substring" - Notion will suggest the matching function

3. Enter your arguments, separated by commas:

  • First argument: The text you want to extract from (e.g., the "Product Name" property)

  • Second argument: The starting point (0 for the first character)

  • Third argument: The ending point (3 to stop before the 4th character)

‍

substring (Product Name, 0,3)

‍

Note: Character positions in Notion start at 0, like in most programming languages. So position 0 is the first character, position 1 is the second character, and so on.

‍

The resulting output will just include the first three letters of the “Product Name” property. 

‍

Converting text to uppercase

With Notion functions, you can also manipulate text cases. For instance, the `upper()` function converts text to uppercase. Let’s see it in action. 

‍

Add an upper() function at the beginning of your formula. The only argument required is the text that you want to capitalize. In our example, we’ll capitalize the substring extracted from Product Name, so we can wrap that entire function in upper’s parentheses.

‍

upper(substring(Product Name, 0, 3))

‍

‍

The resulting output is the same substring as before, but now, all of the text is uppercase. 

‍

Concatenating text

Let's make these IDs a bit more unique by including some text from the page ID field as well. Within the "upper" function, add a plus sign after the substring function to concatenate another piece of text. 

‍

Then, add another substring function - this one will take the last 5 characters of the Page ID, positions 27 to 32.

‍

The final formula will look like this:

‍

upper(substring(Product Name, 0, 3)+substring(Page ID, 27, 32))

‍

‍

The resulting text combines both substrings together directly with no spaces in between. 

‍

‍

If you want to add spaces or other characters between concatenated strings, just enter them as another string in the expression.

‍

For instance, you might want to add a hyphen between the two extracted substrings. 

‍

upper(substring(Product Name, 0, 3)+"-"+substring(Page ID, 27, 32))

‍

‍

The updated formula will add a hyphen between the two substrings, as pictured below. 

‍

‍

Mathematical calculations in Notion formulas

‍

Next, let's take a look at how you can use Notion functions to run mathematical calculations in your formulas. Let's start with some simple addition. 

‍

We'll make a "Total Price" field that adds up the values of the "Price" and "Shipping" properties.

‍

The function for addition is just called add(). Then, each number you want to add up is entered as its own argument, separated by commas. We're just adding up two numbers here, but you can add up as many numbers as you’d like. 

‍

add(Price, Shipping)

‍

‍

The output is the sum of both numbers. 

‍

‍

Note that you can also skip using the function name and just use plus signs instead. You'll still get the exact same result, assuming you're adding numbers together. 

‍

‍

However, if any of your arguments are text strings, then the output will also be a text string and not a number. That's how 3 + 4 will equal 34 instead of 7.

Multiplication

Multiplication works in largely the same fashion. We'll edit the Total Price formula to include sales tax as well. The multiply() function will work well for that. 

‍

You just need to enter two numbers you want to multiply, separated by commas. That will be the product Price multiplied by the sales tax rate.

‍

multiply(Price, Sales Tax) + Shipping + Price

‍

‍

The final output includes the total price with shipping and sales tax.

‍

‍

Using Conditional Logic with an ifs() statement

‍

Of course, sales tax isn't the same in every state in the US. But accounting for variability like that is easy with Notion functions. We just need to use some conditional logic in an ifs() statement. This will let us set the value of a property dynamically based on any condition we want.

‍

We'll edit the Sales Tax field and convert it into a formula field. The function will be ifs(). That will let us specify several conditions and corresponding values.

‍

The pattern of arguments will go like this:

‍

ifs(
{first statement to evaluate}, {value to use if the first statement is true}, 
{second statement to evaluate}, {value to use if the second statement is true}, 
{default value to use if no statement is true}
)

‍

So in our example, we need to start by adding a condition to evaluate. With this double equals sign, we'll check if the "State" field is set to "CT"

‍

ifs(State == “CT”,

‍

Then, we need to provide a value to output if that condition is true. That will be Connecticut's sales tax rate of 6.35%.

‍

ifs(State == “CT”, 0.0635,

‍

In an ifs() function, you can provide as many condition and value pairs as you want. To make it easier to visually parse this function, you can hit shift+return to add new lines. 

‍

Notion formulas ignore spaces and new lines unless they're enclosed in quotes to make a string, so this is purely aesthetic formatting to make things easier to understand.

‍

Once you've added all your specific conditions, you can wrap the function up with a single default value. If none of the other conditions are met, then the function will output this value. As such, it doesn't need a condition.

‍

Here's a complete example, including 0 as the default value.

‍

ifs(
State == "CT", 0.0635,
State == "MA", 0.0625,
State == "ME", 0.055,
State == "RI", 0.07,
State == "VT", 0.06,    
0)

‍

‍

When you save the formula, try changing the “State” dropdown to see the tax and price properties update instantly. 

‍

Working with date/time in Notion formulas

‍

Let's conclude with a brief overview of date/time formulas. 

‍

We'll start with a very easy one: now(). 

‍

‍

With now(), you can retrieve the current time up to the minute. This is another function that doesn't require an argument.

‍

‍

When you enter the `now()` function into a formula, that formula will recalculate the current time whenever the database is reloaded. 

‍

‍

If you want to reference the time the page was created as a fixed value instead, use the "Created Time" property instead. 

‍

‍

That property exists for every page in Notion, even if you haven't manually added a field for it.

Reformatting dates and times with Notion functions

Next, let's create a function that reformats the date generated by now() – or any other date/time value you want to reformat. 

‍

To reformat dates and time, you can use the formatDate() function. The first argument is the date/time you want to reformat - in this case, now(). The second argument is a formatting token enclosed in quotes.

‍

You can check out this fantastic documentation from Thomas Frank for an extensive list of example formatting tokens. 

‍

For example, the format below shows 2 digits for the month and day, and 4 for the year, then shows 12-hour time with AM or PM as appropriate:

‍

formatDate(now(), "MM-DD-YYYY hh:mm A")

‍

‍

Save your formula to see the final output. 

‍

Take your Notion databases to a new level

‍

Notion formulas unlock a totally new level of functionality in Notion. If you need to reformat text, do some math, add conditional logic, or work with dates and times in your databases, then custom formulas are an indispensable tool. The examples we've covered here just scratch the surface of what's possible, but they should give you a solid foundation for creating your own formulas and automating your workflow in Notion. You can explore more options in Notion’s documentation.

If you’d like to learn more about automating your work with Notion, Airtable, Zapier, Pipedream and more, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

AI chatbots have revolutionized how we work with data, but constantly switching between chat windows and your Notion workspace isn't ideal. What if your AI could automatically tap into your company's information and produce results directly alongside your data? That's exactly what Notion's AI autofill properties deliver.

‍

In this guide, we'll explore how to leverage AI autofills in your Notion databases, covering both the built-in prompts and custom options to help you get the most out of this powerful feature.

‍

Getting Started with AI Autofills

‍

To add an AI autofill to your Notion database, simply open your database and create a new property. After the "Suggested" section, you'll find a set of "AI Autofill" options, including three premade choices (summary, translation, and keywords) and the option to write custom autofills.

‍

‍

Keywords: Smart Tagging for Your Content or Data

‍

The keywords option creates a multi-select property with AI autofill enabled. This feature analyzes the content of all other properties in your database to generate relevant tags automatically.

‍

‍

Notably, the AI will reuse relevant existing options in the multi-select field. This will prevent the creation of redundant tags with similar meanings. For instance, in the screenshot below, you can see multiple blogs in our database tagged with “Zapier Automation”. 

‍

‍

To use this feature effectively:

1. Enable "generate new options" to allow AI to create new tags

2. Click "Autofill all pages" to generate tags for every entry

3. Use the magic wand "update" button to generate additional results at any time

‍

Translation: Breaking Language Barriers

‍

The translate option allows you to convert text from one or more properties into different languages. While you can select multiple properties for translation, the results will all appear in a single property, separated by bold headings. To keep your data organized and cleanly structured, we’d generally recommend translating just one property at a time. 

‍

To set up a translation autofill:

• Choose the property you want to translate

• Select a target language

• Generate results

‍

‍

In our tests with the translation autofill, the translations often came across as a bit mechanic and overly literal, but seem to be quite accurate overall. 

‍

‍

Summary: Distilling Information

‍

The summary autofill condenses all properties on a page into a few concise sentences. It's particularly useful for distilling data-heavy entries into actionable takeaways. 

‍

There’s very little to configure with a summary autofill. Your only choice is to decide whether or not the property should autofill whenever a page is edited. 

‍

‍

In our tests, the summary autofill tended to provide generic and inconsistent summaries. For instance, in our blog database, some summaries were written as descriptions of the blog post, beginning with “This blog”, or “This guide”.

‍

‍

However, others were simply shorter versions of the blogs themselves. 

‍

‍

If you want to create more consistent summaries for a specific context, you can use a custom autofill instead. 

‍

Custom Prompts: Tailoring AI Autofills to Your Needs

‍

For more specialized tasks, Notion allows you to create custom AI autofill prompts. Here's how:

‍

1. Create a new property

2. Choose "AI custom autofill"

3. Write your prompt in "What to generate?"

4. Reference specific properties you want the AI to consider

‍

For example, you might create a prompt like: 

‍

"Write a description of the blog to use as an SEO meta tag. Base the summary on the 'Blog body' property. The description should be about 160 characters."

‍

‍

The resulting output creates consistent summaries that are well-suited to being SEO meta descriptions. Each summary is framed with a call to action, and advertises the contents of the blog to the reader. 

‍

‍

Unfortunately, there’s no option to directly insert properties as variables in your prompt. You can find this sort of functionality in Airtable AI, which allows you to reference specific fields in your custom prompts. 

‍

‍

Mentioning specific properties in the prompt seems to work pretty well, but it would be nice to see similar options in Notion to ensure that the AI looks at relevant fields – and ignores the rest. 

Making the Most of AI Autofills

‍

Notion's built-in AI properties streamline the process of interpreting and transforming your company's data. For quick results on common tasks, use the prebuilt options. For specialized needs, craft custom prompts that align with your specific requirements.

‍

Remember that these tools are meant to enhance your workflow, not replace human oversight. Always review AI-generated content to ensure it meets your standards and makes sense in context.

‍

By effectively utilizing AI autofills, you can automate routine tasks, maintain consistency across your database, and free up time for more strategic work. Whether you're managing content, organizing research, or streamlining workflows, Notion's AI capabilities can help you work smarter, not harder.

If you’d like to learn more about streamlining your everyday work with AI and automation, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Are you looking to save hundreds of hours and thousands of dollars for yourself, your company, or your clients? 

‍

We're excited to announce the launch of XRay.Tech's comprehensive Workflow Design Course, designed to teach you how to build powerful, cost-effective automations using popular tools like Zapier, Make, and Airtable.

‍

What You'll Learn

‍

In this fully asynchronous course, we've prepared dozens of video lessons and activities to teach you our battle-tested strategies for workflow automation.

‍

‍

At the core of the course, you'll learn how to build a customizable CRM system from the ground up.

‍

The Power of Custom CRM Development

‍

When we first started XRay, we faced a common challenge: we needed a robust CRM to manage our growing pipeline, but the available options weren't ideal. Off-the-shelf solutions were charging around $1,000 per month and didn't offer the integrations we needed. Instead of compromising, we built our own solution.

‍

Using Airtable as our foundation and automating it with Zapier and Make, we created a custom CRM that has helped us:

‍

• Save thousands of dollars on software costs

• Manage millions of dollars in opportunities

• Create a fully customizable system that grows with our needs

• Automate client onboarding with instant creation of Google Drive folders, Slack channels, and other key assets

‍

Course Highlights

‍

Throughout the course, you'll learn how to build this same cost-effective CRM for your own company, along with the skills to implement your own time-saving automations. Some of the capabilities you'll master include:

‍

• Automatically creating assets

• Sending automated alerts

• Importing external data (e.g., emails) into a database automatically

• Creating an automated activity log

• Triggering automations with webhooks

• Modular, scalable automation design

‍

Real-World Application

‍

At XRay, we've implemented custom workflow automations for hundreds of clients, connecting and optimizing apps like Hubspot, Salesforce, Jira, Asana, Shopify, Magento, and many more. 

‍

Tom Nassr, XRay’s co-founder and CEO, will be your instructor for the course. Tom has been building automated workflows with tools like Zapier for nearly a decade, and has posted over 200 videos about automation and AI to the XRay YouTube channel to help no-code builders learn their craft for free. 

‍

Special Launch Offers

Cyber Monday Promotion

We're offering our lowest-ever price for the course during our Cyber Monday promotion. The first five people to sign up will receive the best possible rate, with prices increasing as more learners join.

‍

Here’s how the promotional pricing will work:

‍

• First 5 seats: $47 (coupon code: xraycourse47)

• Next 5 seats: $97 (coupon code: xraycourse97)

• Next 5 seats: $147 (coupon code: xraycourse147)

• Next 10 seats: $250 (coupon code: xraycourse247)

• Regular retail price: $497

‍

Founding Member Benefits

As the founding members of the course, early enrollees will receive two exclusive bonuses:

‍

1. Access to a private Slack channel for course support and feedback

2. Invitation to a virtual group call where we'll help you customize the CRM for your specific use case

‍

Launch timeline

‍

The presale is live now, with course content becoming accessible starting December 18th. We'll be adding more content after the initial launch, so early enrollees will receive all future updates at no additional cost.

‍

Ready to Transform Your Workflow?

‍

Don't miss out on this opportunity to learn professional workflow automation skills at our special launch price. The first five seats at our lowest price point may be gone by tomorrow!

‍

Sign up now to secure your spot at our lowest available rate.

‍

For brick-and-mortar businesses looking to expand into digital payments, Stripe offers a reliable and trusted solution. While Stripe is incredibly powerful, getting started can feel overwhelming if you're new to digital payment processing. 

‍

Here are five essential tips to help you make the transition smoothly, based on our experience helping traditional businesses embrace digital payments.

‍

1. Prepare Your Business Documentation First

‍

Before diving into Stripe's technical setup, gather all your business documentation. You'll need:

• Your business's legal name (exactly as it appears on IRS documents)

• Employer Identification Number (EIN)

• IRS letter 147C or similar verification document

• Business bank account information

• Business owner's Social Security number and other personal info

‍

‍

Pro tip: If you need a 147C letter from the IRS, call them directly to request it. While they may not explicitly offer a "147C letter," they will send you official documentation containing your EIN and legal business name, which Stripe will accept for verification.

‍

2. Understand the Difference Between Products, Prices, and Payment Links

‍

Stripe organizes your offerings in a hierarchy:

‍

• Products are the base items or services you're selling

• Each product can have multiple prices (perfect for different tiers or variations)

• Payment links connect your products to customers without requiring any code

‍

Start by creating your products first, then set up your prices, and finally create payment links. 

‍

‍

This structured approach will make it easier to manage your digital offerings, especially if you're converting physical goods or services into digital payments.

‍

3. Take Advantage of No-code Payment Links

‍

Payment links are your best friend as a no-code business owner. They allow you to:

‍

• Accept payments without writing any code

• Collect customer information like shipping addresses and phone numbers

• Add custom fields for business-specific needs (like license plate numbers for parking services)

• Enable multiple payment methods including Apple Pay and Google Pay

• Implement promotional codes and discounts

‍

‍

The best part? You can easily embed these payment links into any website builder platform, from WordPress and Wix to Squarespace and Softr. They’ll also work great on XRay Workflow. 

‍

To create a payment link, just select “Payment links” under the “Shortcuts” menu, and click “New”. 

‍

‍

4. Set Up a Testing System Before Going Live

‍

Before accepting real payments, set up a testing system to safely verify that everything works correctly:

‍

1. Create a 100% off coupon code for internal testing

2. Use this code to run complete transaction tests

3. Verify that customer information is collected correctly

4. Check that confirmation emails and receipts work as expected

5. Test the customer experience from start to finish

‍

‍

This approach lets you validate your setup without risking real money or creating confusion in your accounting.

‍

To create a new coupon, open up your product catalog, select the “Coupons” tab, and click “New”. 

‍

‍

5. Consider Additional Features for Better Customer Experience

‍

Once your basic setup is running, consider enabling these helpful features:

‍

• Two-factor authentication for account security

• Automatic tax calculations based on your location and product type

• Climate contribution options (contributes 0.5-1.5% of your transactions to C02 removal)

• Custom receipt and invoice generation

• Address collection for customer records

‍

‍

These features can help create a more professional and trustworthy experience for your customers while making your business operations more efficient.

‍

Start collecting payments with a hassle-free system

While transitioning from brick-and-mortar to digital payments might seem daunting, Stripe's no-code options make it entirely manageable. The initial setup might take some time – our recent parking lot client took about a month to complete verification – but the result is a professional, reliable payment system that can help grow your business.

‍

Remember, you don't need to be a technical expert to accept digital payments. With these tips and Stripe's payment links, you can create a professional digital payment system that works alongside your existing business operations.

If you’d like to learn more about optimizing and automating your business, be sure to check out the other posts on check out our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Notion is a great app for organizing your team’s data and documentation, but compared to a platform like Airtable, it has always been rather lacking when it comes to automation. 

‍

And while they still haven’t closed that gap, Notion is making some impressive progress with their internal automations. 

‍

In this post, we’re going to show you how to use the latest database automation features in Notion to help you save time and get more out of your Notion workspace. 

‍

We’ll start with a very quick Slack automation that you can build with a free Notion plan. Then, we’ll show you some of the more advanced features that require a Notion Plus account. 

What you’ll need to follow along with this tutorial

• Notion - free account (for the first automation)
• Notion Plus account (for the second automation)
• Slack - free account
• GMail - free account

Building a simple automation: Slack alerts for new pages in a Notion database

‍

To begin, we’re going to build an easy automation that just sends a Slack alert whenever a new page is added to a database in Notion. 

‍

This is an automation that you can build with free plans in both Notion and Slack. 

‍

Summary of steps:

Create a new automation

Choose a trigger: “Page Added”

Pick an action: “Send Slack notification to…”

Connect your Slack account to Notion

Select a Slack channel

Test the automation

‍

Creating a new automation

To get started with database automations in Notion, just open up the database you want to automate. Then, click on the lightning bolt up in the top right next to the search icon. 

‍

‍

This will immediately create a blank automation canvas that you can configure with custom settings. 

‍

‍

Under “When”, you’ll need to choose a trigger. In workflow automation, a “trigger” is an event that prompts an automation to run. 

‍

For this workflow, pick “Page added”. That means this automation will run whenever a new page is added to this database. 

‍

‍

If you’d like to restrict the automation so that it only runs for new pages that appear in a specific view, you can select that here where it says “for all pages in”. 

‍

‍

However, for this example, we’ll have the automation watch the entire database for new pages. 

‍

Adding an automated action: Send a Slack message

Under “Do”, you can pick an automated action to perform.

‍

On the free plan, you’ll see that only one option is available: “Send slack notification to…”

‍

‍

This will automatically send a precomposed Slack message to a channel of your choice whenever your trigger event occurs. 

‍

There’s not much to configure in this action. The only setting you can customize is the channel that the message will be sent to. 

Connecting Slack and Notion

When you click to select a channel, you’ll be prompted to connect a Slack account to Notion if you haven’t done so already. Click “Connect” to add a Slack account. 

‍

‍

You’ll then see two pop-up windows appear in succession. The first pop-up will let Notion access Slack. 

‍

‍

And the second one will let Slack access Notion. 

‍

‍

Choose a channel for your notification

Now that Slack and Notion are both authorized, you can pick the channel you want to use. 

For our example, we’ll choose a “tutorials” channel. 

‍

‍

For this action, there’s nothing else to configure, so you can proceed to testing the automation. 

Testing your automation: using forms to add pages to a Notion database

‍

Whenever you create an automation that runs for new pages in a Notion database, we’d strongly recommend using forms to add new pages rather than adding them directly to the database. 

‍

If you create a new blank page and then fill in the info one field at a time, the automation might run before you can finish providing the necessary information. Consequently, the automation might not run correctly, as it could be missing critical information. 

‍

You can click on the plus button to create a new view, and select “Form”. 

‍

‍

This will automatically create a form for every editable field. 

‍

‍

Under “Share”, copy the link for your form and open it in a tab. 

‍

‍

Fill out the form to create a new page with all of its information already in place. 

‍

‍

Notion automations run every few minutes, so the Slack message won’t appear right away. However, within a couple minutes, you should see the Slack alert with Notion’s preconfigured message, which includes links to the database and the new page. 

‍

‍

This automation is all set. If you want to edit it or turn it off, you can find those options by clicking on the automation icon, and selecting the three-dot icon next to the automation’s name.

‍

Expanding your database automation with Notion Plus

Next, let’s take a look at what Notion Plus unlocks for database automation. 

‍

We’ll build a similar but more advanced automation. It will still trigger when a new page is created in this database, but instead of sending a Slack message, we’ll have it compose an email in Gmail. 

‍

‍

We’ll also include a custom variable that uses a formula to dynamically retrieve the URL of the new page. 

‍

Create a new automation and choose a trigger

To begin, open up a database, and click on the lightning bolt automation. Create a new automation, and choose “Page added” as the trigger again. 

‍

Add an action: Send mail

For the action, you’ll see that there are now several more accessible choices. you can edit properties in a database, create new pages, edit pages, send Notion notifications, or define custom variables.

‍

‍

For this first action, choose “Send mail to…”. 

‍

Just like with Slack, you’ll need to connect a Gmail account before you can complete your setup. 

‍

‍

‍

Now that Notion is authorized to use Gmail, you can compose the message you want to send. 

‍

‍

For the recipient, you can either choose to send the email to a specific account attached to your Notion workspace, or dynamically pick an address from an email field in the table. 

‍

In our example, we’ll just send all of these alerts to our XRayTutorials email. 

‍

‍

For the subject line and message, you can enter static text directly, and insert dynamic data retrieved in the trigger. 

‍

In our example, we’ll begin typing a simple subject line - “A new contact was added to the Directory database on…”

‍

‍

Then, we’ll insert some dynamic data for the date and time. You can type @ to pull up data retrieved in the trigger. 

‍

‍

You’ll see all of the page’s attributes, as well as some metadata like the date and time the automation was triggered. 

‍

For our example, use “time triggered” to display the date and time the new page was created. 

‍

‍

In the message, we’ll provide some more detail about the page, like its title, and the email address of the new contact that the page represents. We’ll also include the URL of the page. 

‍

Add an action: using a formula to update a property with conditional logic

The Gmail step is all set, but let’s add one more step here before we finish and test the automation. 

‍

Let’s use Notion’s formulas with the “Edit property” step to update a field to a different value based on the value in another field. In our example, we’ll change the contents of the “Language” field based on whatever’s in the “Country” field. 

‍

Add a new step at the end of the automation. Choose “Edit property”. 

‍

‍

Choose the field you want to edit. For instance, we’ll be updating the “language” field. 

‍

Then, click on the sigma symbol (Σ) to write a formula. 

‍

‍

We’ll start with an ifs statement, because we want to change the value of this variable based on different conditions. 

‍

First, we need to specify a condition to evaluate. We’ll check if the “Country” field equals “UK”. You can see the syntax for that condition in the screenshot below.

‍

‍

Then, we'll provide the value to use for the “language” attribute if that condition is true – that is, if the country is UK. We’ll enter “English”. 

‍

‍

You can add as many conditions as you’d like to an ifs statement. We’ll add one more for “France” and “French” using the same syntax. 

‍

‍

Finally, we’ll wrap it up with a default choice - “No language provided”. 

‍

‍

If none of the other conditions are true, then Notion will use this default choice, so we don’t need to provide a specific condition to test – just the value to use. 

Learning more about formulas in Notion

Formulas can do a lot to expand the functionality of Notion automations, so we’d recommend checking out Notion’s docs to learn more about how they work. 

Testing the finished automation

This formula’s all set, so we’ll click on “Save” to save the step. 

‍

‍

Then, we’ll click “Create” to complete the automation. 

‍

‍

Once your automation is finished, give it a few tests to make sure that conditional logic in the formula works as intended.  

‍

For instance, we’ll submit three form responses: one from the UK, one from France, and one from India, which should trigger the default condition. 

‍

‍

After waiting a moment, you should see some email alerts for each new contact. 

‍

‍

And if you check the database, you should see that their languages are all configured correctly. 

‍

Improved automation features for Notion power users

Ultimately, Notion’s automation tools are still not much to boast about in contrast to Airtable – and they certainly can’t compete with dedicated automation providers like Zapier, Make, or Pipedream. 

‍

If you’d like to take your no-code automation game to the next level, we’ve prepared a beginner’s guides for all three:

‍

Pipedream beginner’s guide

Zapier beginner’s guide

Make beginner’s guide

‍

However, for Notion power users or organizations who are already storing their data in Notion, these updated automation features will be a quick and convenient way to automate simple tasks in an app you’re already familiar with. 

If you’d like to learn more about automating your work with the software you use every day, be sure to check out the other check out our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Are you a software developer or a low-code builder looking for a fast and functional way to build automations?

‍

While providers like Zapier and Make are great for connecting supported apps together with a simple visual interface, their options for adding your own custom code are pretty limited. 

‍

If you’d like to use an automation provider that gives you the freedom to enhance your automations with custom code, sync your projects to Github, and easily write any API call you’d like, then you should explore Pipedream. 

‍

In this comprehensive guide, we’re going to show you everything you need to know to get started. 

‍

We’ll begin by explaining what Pipedream is, and showing how its unique features stack up to competitors like Zapier and Make. 

‍

Then, we’ll show you how to build your first automated workflow in Pipedream. After we’ve built the basic, 2-step workflow, we’ll give it some enhancements with a formatting step, some custom code, and conditional logic. 

What is Pipedream?

Pipedream is an advanced automation provider that lets you build automated workflows by connecting web apps together.

‍

When a triggering event occurs in one web app, the automation will start running, and will perform actions in one or more other web apps. 

‍

‍

For seasoned no-code builders, this probably sounds very similar to Zapier or Make. In many ways, the comparison is appropriate, as you can certainly use Pipedream as a no-code platform if you’d like. They have pre-built integrations for over 2,000 apps and counting, and you can automate your work with these integrations without needing to write any code at all. 

‍

‍

Advanced technical features

However, Pipedream includes many technical features that distinguish it from its competition. On their site, they promise “code-level control when you need it and no-code when you don’t”. 

‍

‍

In our experience using the app, that claim rings true. 

‍

Pipedream gives you the speed and convenience of a no-code platform, with everything being deployed and hosted on Pipedream’s servers.

‍

However, it also gives you the flexibility of adding custom code when you need it, and their options for custom code go far beyond providing a plain text editor to paste in a function or two. 

‍

Extensive support for custom code and engineering best practices

The code editor in Pipedream features syntax highlighting and suggestions, so you can write JavaScript, Python, Golang or Bash with many of the same features you’d expect from your favorite IDE. 

‍

‍

On top of that, you can also sync your Pipedream projects to Github for robust version control. This is a really nice feature for keeping track of your changes. 

‍

‍

Additionally, if an app that you want to automate isn’t natively supported in Pipedream, you can always write custom API calls to perform the actions your workflows need. 

‍

‍

Features like environment variables make it easy to securely use API keys for authorization when you write your API requests.

‍

Customizable timeout settings

Another unique and extremely useful feature in Pipedream is the ability to adjust the timeout window for each automation. 

‍

By default, each workflow will wait up to thirty seconds to run each step, but you can raise that all the way to 12 and a half minutes on a premium plan. 

‍

‍

Your workflows will consume more credits if they use a longer timeout window, but this can be an indispensable feature if your automation relies on lengthy actions. 

‍

For instance, you might need to wait a while for an AI to process a long prompt, and there’s simply no way to do that on Zapier or Make, which have hard timeout limits that you can’t adjust. 

An ideal platform for low-code engineers

If you’re a software developer or low-code engineer looking for a fast way to build an automation without compromising on functionality, Pipedream is a great pick. 

‍

It’s the perfect tool for connecting your internal or legacy systems to modern, commercially available web apps. 

‍

Pipedream is also an ideal choice for low-code builders who just want more control over their automations. 

‍

Additionally, for anyone looking to build a HIPAA-compliant workflow automation, Pipedream is an option. They will sign a BAA, provided that you speak with their team and explain your use case. 

‍

Building your first automated workflow in Pipedream

Now that we’ve covered the basics, we’re going to show you how to build your first automated workflow in Pipedream.

‍

This process will be very straightforward to follow for everyone, regardless of your technical background. 

‍

To begin, we’ll create a very simple automation. When a custom label is added to an email in Gmail, the contents of that email will be automatically sent to Slack as a channel message. 

‍

‍

Once we’ve built that initial automation, we’ll add some additional features, like conditional logic to automatically summarize long emails with OpenAI. 

‍

‍

However, we want to note that the focus of this tutorial is not on these specific apps. Instead, our goal is to show you how Pipedream works using this software as an example. Automating other apps will follow very similar principles. 

Follow along for free

You can follow along with us on a free Pipedream plan. You will need a paid plan to deploy the finished workflow, but you can still build and test it without starting a subscription. 

‍

There’s also a free trial available if you’d like to explore Pipedream’s premium features. 

Create a new project and workflow

To begin, sign in to Pipedream, and create a new Project. Give your project any name you’d like. 

‍

‍

In Pipedream, every workflow needs to belong to a project, which can include multiple workflows, as well as file stores and variables. 

‍

‍

In the “Resources” tab, click “new” to create a new item, and select “Workflow” to make a new automation. Be sure to give your workflow a descriptive name.

‍

For the sake of this tutorial, you can leave all of the other settings to their defaults and just click “Create Workflow” to get started. 

‍

Add a trigger to your workflow: Watching for newly labeled emails in GMail

Every workflow in Pipedream needs to begin with a trigger. Click “Add trigger”, then choose an app where your trigger will occur. 

‍

‍

For our example, that will be Gmail. 

‍

‍

Then, choose the specific triggering event you want. That will be “New labeled email”, so the automation will run whenever a label is added to an email. If you’re not sure which event to choose when building your automations, just check the provided descriptions for more detail. 

‍

‍

Authorize the trigger app

Next, you’ll need to authorize the app used in your trigger. The authorization requirements will vary based on the app in question. 

‍

For Gmail, you’ll just need to sign in with a simple OAuth portal. 

‍

‍

In other apps, like OpenAI for instance, you’ll need to provide an API key. 

‍

Configure your trigger settings

Once you’ve authorized your trigger app, you can configure your trigger’s settings. 

‍

Different triggers will have different settings, but they’ll typically let you set filters and limits on what data to watch for. 

‍

‍

For this “New labeled email” trigger, you can choose a specific label that Pipedream will watch for. The workflow will only run when that specific label is applied to an email. 

‍

For our example, create a custom label in GMail named something like “Send to Slack (Pipedream)” label. 

‍

‍

Be sure to apply the label to at least one email in your inbox, as well. This will be used as test data soon. 

‍

Return to Pipedream, and select the “Send to Slack”  label in your GMail trigger. You may need to refresh the data first by clicking “Load More” to see your newly created label. 

‍

‍

In Pipedream triggers, you can also change the “Timer” setting to choose how often your workflow will search for new data. Note that some shorter intervals will require a paid plan. 

‍

‍

Once you’ve configured your trigger, click “Save and continue”. 

Testing the trigger

With your trigger configured, you’ll need to load some test data into your workflow. 

‍

You can either use real data retrieved from your trigger app, or automatically generate mock data with Pipedream. 

‍

‍

We’d generally recommend creating your own test data so that you can control exactly how it’s set up. To create test data for our example workflow, just apply the “Send to Slack” label to an email in your GMail inbox if you haven’t already. 

‍

If Pipedream has found a relevant piece of data, you’ll see an alert about a new “Event”. 

‍

‍

Click on the dropdown to see more details about the event and select it as your test data. 

‍

‍

After a successful test, you can see all of the data that the trigger retrieved for use in your automation under the “Exports” tab

‍

‍

Clicking on the toggles will display additional information. For instance, under “decodedContent”, you can see the email’s body. 

‍

‍

And under “parsedHeaders”, you can find the email address and name of the sender, among other data.

‍

‍

Add an action to your workflow: Sending a message in Slack

With a properly configured trigger and test data in place, you can now add additional actions to send this data to any other app you’d like. 

For our example, we’ll take this email data and send it to Slack. 

‍

Click “Continue”, and choose an app for your next action. Search for the app you want to use (in this case, Slack). 

‍

‍

Next, choose an action. 

‍

Note that if you don’t see the action you want to perform in your chosen app, you can always choose to build a custom API request instead, or use AI to generate a custom action. 

‍

But that won’t be necessary for this workflow, since you can just use the pre-built “Send Message” action. 

‍

Authorize the app used in your automated action

After choosing an app and event, you’ll need to authorize the app just like you did for the trigger app. 

‍

For Slack, that will just require signing in to your Slack account, and clicking “Allow” to grant the appropriate permissions.

‍

Configure your action

Just like with triggers, the settings for any action in Pipedream will depend on the app and action in question. 

‍

But typically, you’ll need to specify where you want to create or update data, and what the data is that you want to create or edit. 

‍

If you’re not sure what the settings are for, or what your options are, you can just read the provided help text (highlighted in the image below). 

‍

‍

For Slack, you need to pick the channel type – direct message, group, private channel or public channel. In our example, we’ll be sending a channel message. You may prefer to send a direct message to make your automation a bit less “noisy”. 

‍

‍

Then, you need to specify a channel, or direct message recipient. For example, we’ll send our message to a dedicated “tutorials” channel. 

‍

Using dynamic data retrieved from earlier steps

Finally, you need to compose your message in the “text” field. In most data fields in Pipedream, you can enter text directly, and you can provide dynamic data retrieved from previous steps in the workflow. Try using both dynamic data and directly entered text in your message. 

‍

You can access dynamic data from the window that pops up. The data all be sorted under “Event”, and other headings specific to each app. 

‍

‍

For instance, you can find the sender’s email address from the GMail trigger under the “parsedHeaders” toggle. 

‍

‍

Just click on “Select path” to insert any variable into a field. 

‍

‍

By inserting the “from.email” path into your message, you’ll ensure that the Slack message displays an address retrieved from the email that triggered the message in the first place – not just a static email address. 

‍

In our example, we’ll also include the send date and time, along with the email body - that’s the “decodedContent” variable. 

‍

Optional settings

That takes care of all the required fields for this action, but there are also a few optional settings that are worth checking out. 

‍

For instance, with these optional settings, you can configure the workflow to use a custom bot name and emoji. We’ll call ours “Email Assistant Émile”, and use the :email: emoji. 

‍

‍

Additionally, you can also choose whether or not to “unfurl” links. We usually set this to “FALSE” to make sure that links don’t take up too much space in messages.

‍

Testing the automated action

Once your message is fully configured to your liking, click on “Test” to run your action. In this case, that will send a message in Slack. 

‍

‍

After a successful test, you’ll see details about the message sent in the “exports” tab. 

‍

‍

You can open up your Slack to see the message that was sent. 

‍

‍

Your first automated workflow in Pipedream is all set, and you can click on “Deploy” to start using it. 

‍

‍

Once it’s deployed, the automation will run on your chosen schedule to check for new emails with your custom label. 

Enhancing your workflow with formatting, custom code, and conditional logic

The automation that we’ve just built works fine, but it’s a very basic setup. 

‍

There are a few things we can do to make it better, and these quick improvements will also introduce you to a few key features and concepts in Pipedream. 

‍

In this section, we’ll explain how to add formatting, custom code, and conditional logic to your workflow. 

Formatting a date/time value

The formatting of the date and time in our example Slack message is a bit awkward and overly long. In a context like this where we’re optimizing for human readability, we don’t really need a timestamp that includes seconds and the timezone. 

‍

Thankfully, Pipedream has built-in features for easy data formatting. 

‍

Pause the deployed workflow so it doesn’t run while you’re making edits, and start editing it. 

‍

‍

‍

After the trigger, add a new action, search “format” and choose the “formatting” app. 

‍

‍

This is a built-in utility created by Pipedream. It includes several options for converting to JSON, comparing dates, extracting URLs and much more. For this automation, choose “Date/Time format”. 

‍

‍

With this action, you can just enter the date/time value you want to format. Then, you can choose the format you want to use instead. If you’re not sure of the syntax to provide, just click on the link in the help text for more info. 

‍

In our example, we’ll use the {long} preset, which will display the date like this:

‍

October 6, 2024 2:38 PM

‍

‍

Give your formatter step a test to make sure the output matches your desired format. 

‍

‍

Finally, make sure to update the Slack message to use the reformatted date/time value instead of the original. 

‍

Conditional logic and custom code

Next, while the Slack message looks good overall, this particular email is awfully long. 

‍

Let’s update this automation so it will give us an OpenAI summary of any email that’s longer than 1,000 characters.

To begin, we’ll need a way to determine the length of each email. 

‍

There are two ways you can do this - either with custom code, or with an expression built into the conditional logic. 

Finding the length of a text string with custom code

Even though it’s not the ideal way to do it, we’ll begin with custom code, since this will be a good opportunity to demonstrate how custom code works in Pipedream. 

‍

After the formatting step, add another action, and choose “Run custom code”. 

‍

‍

The custom code step will default to JavaScript, but you can also choose Python, Golang, or Bash. 

‍

‍

If you have some experience with web development, this custom code module will seem pretty straightforward and familiar. 

‍

If you’re not experienced with JavaScript, just know that the “run steps” line is importing all of the data from the previous steps in the automation so you can easily reference it in your code. 

‍

‍

Add one line to create a variable called “bodyLength”, and set it to be equal to the character length of the email body:

‍

let bodyLength = steps.trigger.event.decodedContent.length

‍

Then, add another line to return the value of that variable in this step’s output:

‍

return bodyLength

‍

You can refer to the screenshot below to see the full custom code that we’re adding. 

‍

‍

Note that helpful suggestions come up as you type in the custom code editor, making it easy to find what you’re looking for. 

‍

Give this step a test, and you’ll see the returned character length. In our example, we can see that our email is definitely over 1,000 characters. 

‍

Adding conditional logic to a workflow

Now let’s add conditional logic to evaluate that number. 

‍

Add a new action, and under “control flow” choose “If/else”. 

‍

‍

First, you’ll need to specify the condition to evaluate. Use the condition pictured below to check if the bodyLength variable is greater than 1,000. 

‍

When you test this condition, you’ll see whether or not the processed data would pass the filter. 

‍

‍

In our example, the body length is greater than 1,000, so the automation would continue along the first path on the left. 

‍

‍

If the number was not greater than 1,000, then it would continue along the second path on the right. 

Using an inline function to calculate the length of a string

You can also insert the same conditional logic without using a custom code step.

‍

Instead, you can just enter the “decodedContent” path in the condition itself, then append the .length function before the curly brackets. 

‍

‍

Most simple actions can be performed in Pipedream without code, so you’ll likely only use custom code for more complex tasks, like array comparisons, or complicated loops.

‍

However, we still wanted to give you a brief example of how it works to get a little practice with it. 

‍

Now, let’s quickly finish up this workflow. 

Adding an OpenAI step to the workflow

In the first path for data that passes our condition – in other words, for long emails – we’ll add a ChatGPT step that sends a prompt asking to summarize the email. 

‍

You can see the final configuration of this step in the screenshot below. 

‍

‍

Note that you’ll need to be set up with an OpenAI API billing plan to use a ChatGPT step like this. You can learn more about using ChatGPT and OpenAI with no-code automation in this tutorial. 

‍

Give the ChatGPT step a test, and check the output. 

‍

‍

Copy and paste the Slack message action into this path, and update it to include ChatGPT’s summary instead of the original message. 

‍

‍

Finally, you can move the original Slack message into the second path for shorter emails. 

‍

‍

In its updated configuration, this automation will run every 10 minutes to check for emails with the “Send to Slack” label. 

‍

If the email’s body is longer than 1,000 characters, it will get an AI summary from ChatGPT and send us the summary in Slack. 

‍

If it’s a shorter email, then it will just send us a Slack message with the original email body. 

‍

With these changes, this automated workflow is all set. 

‍

You can click deploy to commit your changes, but since this workflow uses some conditional logic, you will need a paid plan to complete deployment. 

‍

‍

If you haven’t yet, it's time to kick off that free trial!

Pipedream pricing

While we’re on the subject of a premium Pipedream plan, let’s wrap up this beginner’s guide with a quick look at Pipedream’s pricing. 

Credits in Pipedream

First, we’ll explore how credits work, since they’re a key part of Pipedream’s pricing scheme. 

‍

Your workflows’ data and bandwidth usage is measured in “credits”. 

‍

When you create a new workflow, it will use a 30-second timeout window and will include 256 MB of memory. With these settings, your workflow will consume 1 credit when it runs, regardless of how many steps are in it. 

‍

Longer timeout windows and more memory will require more credits. You can see how many will be used when you adjust the settings. 

‍

‍

This is quite a bit different from Zapier and Make, which charge you for a “task” or “operation” for most steps in your workflow.

‍

Both models have their pros and cons. With Pipedream’s setup, you’re not punished for making long, complex automations. However, with Zapier’s approach, you can process an unlimited amount of data. The better choice will depend on your specific use cases and priorities. 

‍

Pricing plans

Now that you understand how credits work, let’s take a look at the available pricing plans. You can also refer to Pipedream’s pricing page for more info. 

Free plan

The free plan will give you everything you need to try out Pipedream and build a few workflows. 

‍

However, you’ll be limited to three active workflows, 3 connected accounts, and just 10 credits per day. 

‍

Like with most automation providers, the free plan won’t be a suitable long-term choice for business-critical automations. 

Basic plan

For $45 per month, the basic plan will let you build 10 active workflows and will provide you with most of Pipedream’s core features, along with 2,000 credits each month.

Advanced plan

However, you’ll need the Advanced Plan at $75/mo to unlock “premium” apps like Jira, Shopify, Stripe, and Zendesk. 

‍

You can find a full list of Premium apps here. 

‍

The advanced plan also includes conditional logic like the if/else action we created in this tutorial, as well as unlimited active workflows. 

‍

For both the basic and Advanced plans, you can get a considerable discount by paying on an annual basis instead of monthly. 

Business plan

The business plan enables syncing with Github, as well as a long list of advanced technical options. You’ll need to contact Pipedream’s sales team to get a quote for a business plan. 

Fees for credit overages

If you ever exceed your monthly allotment of credits, you’ll just need to pay an additional fee based on how many credits you use.

‍

The more credits you need, the better your rate will be. 

‍

So you don’t really need to worry about your automations stopping suddenly if they’re running a lot in a given month. However, if you want to stick to a specific budget, you can enforce a credit limit in your account’s billing settings. 

Pricing compared to Zapier

Pipedream’s pricing is essentially comparable to or better than Zapier. 

‍

The basic plan is a bit more expensive than a similar Zapier option, but an Advanced Pipedream plan is $30 a month less than a Zapier team plan. 

‍

‍

Beyond their base pricing, Pipedream offers additional credits at a more affordable rate when compared to Zapier’s tasks. 

‍

Adding 100,000 tasks to a Zapier team plan costs $600 a month, compared to $450 a month for the same number of credits in Pipedream. And that gap only gets wider as you purchase a higher volume of credits. 

‍

It’s also worth remembering that a task in Zapier is just one automated step, while a credit in Pipedream can potentially represent an entire workflow, depending on your timeout and memory settings. 

‍

So if you’re automating at a large scale and using millions of tasks or credits each month, Pipedream will be the more affordable choice. 

Build low-code automations with all the technical features you need

With Pipedream, building flexible low-code automations is fast and easy. And for more technical users adding custom code to extend the functionality of your workflows is just as simple. 

‍

If you’re looking for implementation support to get your team started with Pipedream, just schedule a free discovery call right here on our website. We can help you to automate and streamline even the most complicated workflows. 

If you’d like to learn more about automating your work with software like Pipedream, check out the other tutorials on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Loops are one of the most powerful tools available to low-code builders. They allow you to process every item in a dataset individually—whether you're working with 5 items or 500—without adding redundant actions to your automation. In this guide, we'll walk through everything you need to know about creating and using loops in Zapier.

‍

Prerequisites

‍

Before we dive in, note that you'll need:

• A Zapier Pro plan or higher to publish Zaps with loops

• At least one trigger step in your Zap (loops can't be used as triggers)

• Optional: Any additional actions or search steps to collect data for your loop

‍

Creating a loop from line items

Summary of steps

‍

1. Add a new action to your Zap

2. Select "Looping by Zapier"

3. Choose "Create Loop from Line Items"

4. Add your value sets:

   • Select the fields you want to loop through

   • Give each value set a descriptive title

   • Add as many value sets as needed for your automation

5. Optional: Set a limit for iterations (maximum 500)

6. Test your loop to preview the structure

‍

Adding a loop to your Zap

To add a loop to your Zap, add a new action, and select “Looping by Zapier”. 

‍

‍

Choose how you want to create the loop. You can choose to create a loop from line items, text, or numbers.

‍

‍

We’ll explain how text and number loops work later in this tutorial, but we’ll begin with a line item loop. 

Understanding Line Items

In Zapier (and JSON), line items are objects that contain arrays (lists) of other objects. You'll find line items in many common applications:

• Shopping cart items in Shopify orders

• Database records with multi-select fields

• Linked records in Airtable

‍

Many datasets that you work with will be formatted with line items already, but you can also convert data into line items if needed. You can learn more about converting data into line items in Zapier’s help docs. 

Add value sets to your loop

Whatever you choose as your data source, you’ll need to provide at least one value set to loop through. A “value set” represents a field that your loop will process. 

‍

Click on “Add value set” to add a new set to your loop. 

‍

In our example, we’ll create value sets for the “Name”, “Task Count”, and “Slack ID” fields. All of these fields will be retrieved from the Airtable search step that precedes our loop. 

‍

‍

You can also set a limit for how many iterations your loop should run through. Each loop can run up to 500 times, but you can also set a lower limit if you’d like. 

‍

‍

Once you’ve added all the value sets that you want to loop through, and configured your other settings, click continue. Then, test your loop. 

‍

In the test results, under “prevew_loop_values”, you’ll see the structure of your loop. 

‍

‍

Each iteration will include one instance of each value from your value sets. So in our example, each of our loop iterations includes a name, a task count, and a Slack ID, since those are the three value sets we created. 

‍

Each iteration also includes an iteration number by default. 

‍

The data that you see in the loop preview will be all of the data that your loop will retrieve and process as it runs, so make sure that you see the data you want to use in your loop before continuing. 

‍

Once your step is all set up, it’s time to add actions to your loop.

‍

Adding actions within your loop

‍

Any action that you add after a loop step will be repeated once every time the loop iterates (unless the iteration stops at a filter first - we’ll cover filters in the next section). 

‍

Every action before the loop will only run once, just like actions in any other Zap. In our example, the search step that we run after the trigger isn’t part of the loop, so it will only run once each time the automation runs. 

‍

‍

A rectangle starting at the loop step will give you an easy visual reminder of which actions are included in your loop. 

‍

In our example, we’re going to use Slack to send a direct message to each person referenced in the loop. In the Slack message, we’ll include a personalized greeting for each user and a notice with the number of tasks that are currently assigned to them. 

‍

‍

When you’re using dynamic data from a loop, be careful about which variables you pick. The variables that start with “Preview Loop Values” will include data from several iterations. 

‍

‍

If you want to reference the data that belongs to a specific iteration, just use the variables that don’t start with “Preview Loop Values”. For instance, we’ve used the “Name” and “Task Count” variables in this Slack message. 

‍

‍

Testing a looped action

Once you’ve finished configuring your action, give it a test. 

‍

‍

When you test an action in a Zapier loop, it will only process test data for the first iteration in the loop. It’s not going to run test actions for every item. 

‍

So even though our first step retrieved three records to loop through, testing the Slack step in the loop only sent one message, which is what we expected. 

‍

‍

Using a filter to add actions after your loop

‍

Every action that you add after your loop step will be included in your loop. However, if you’d like certain actions to only run once after your loop, you can add a filter step. 

‍

By configuring the filter so that only the final iteration passes, you can ensure that any actions added after the filter only run once, even though they’ll still technically be considered part of the loop. 

‍

To add actions to your loop that only run once:

‍

1. Add a "Filter by Zapier" action

‍

‍

2. Set the field to "Loop Iteration is Last", and set the condition to "(Boolean) Is True"

‍

‍

3. Add any actions that should run after the loop completes.

‍

This setup ensures your post-loop actions only execute once, regardless of how many iterations your loop processes.

‍

In our example, we’ll add another Slack action that sends a channel message. The message will include a list of everyone who received a direct message about their tasks. 

‍

‍

Running a live test of your loop

To see your full loop in action, publish the Zap and turn it on. 

‍

‍

Then, fulfill your Zap’s trigger condition. For our example, we just need to wait until the scheduled time in our trigger, which we’ve adjusted to occur in a few minutes. 

‍

When your Zap runs, the loop will iterate for each item in your chosen data set. Every action in the loop will run once per iteration, except for any actions that you added after the filter. 

‍

When our example automation runs, the Airtable search step retrieves three records in the “People” table. Each of the three people is sent a direct message. 

‍

‍

On the final iteration, the filter is cleared, and a single Slack channel message is sent summarizing the people who received messages. 

‍

‍

Alternative loop types: text loops and number loops

Text loops

A text loop is very similar to a line item loop. However, instead of iterating through several pieces of structured data, it will loop through pieces of text. 

‍

For the loop to work correctly, your text will need to contain a “delimiter” between each item – like a comma, a semicolon, or even a space if you’d like (written as [:space:] )

‍

‍

It doesn’t matter what the delimiter is, as long as the delimiter is used consistently in the text. For instance, you can’t have some items separated by a comma and others separated by a semicolon. 

‍

‍

Each iteration of the loop will process a piece of text until it hits a delimiter. If you were to loop through the text “Apples, Pears, Bananas” using a comma as a delimiter, the loop will iterate three times: once for Apples, once for Pears, and once for Bananas. 

‍

Number loops

‍

When you create a loop from numbers, you won’t be iterating on data from previous steps in the Zap. Instead, you’ll be incrementing a number in a set pattern. 

‍

You can set how many times the loop will run, as well as the loop interval number. So if you’d like, your loop iteration number can increase by 5 or 10 each time instead of just 1. 

‍

For instance, in the loop pictured below, the iteration counter will start at zero and increase by 10 for every iteration until hitting the counter limit of 100. 

‍

‍

When it runs, the loop will just return its iteration number and repeat the steps you add to the loop. However, that iteration number is all the data the loop itself will process. 

‍

‍

Number loops are useful if you want to perform the exact same action multiple times in your Zap, or if you want to use the loop as part of a calculation. 

‍

But you’ll probably be using a line item or text loop in most cases. 

‍

Loops: Frequently Asked Questions

‍

Q: Can you nest loops within loops?

A: No, Zapier doesn't support nested loops directly. Alternative solutions include:

• Using custom code steps

• Connecting separate Zaps using webhook triggers

• Refining your trigger to require only one dataset

‍

‍

Q: How does task usage work with loops?

A: The loop creation step doesn't consume a task. However, every action within the loop consumes tasks normally on every iteration. 

‍

A: Can you use loops with paths?

Yes, but there are some limitations. You can only use one loop per Zap, so you can’t add loops to multiple paths in a single Zap. 

‍

However, you can add paths to a loop as pictured below, which will still enable you to add quite a lot of conditional logic to your automations with loops. 

‍

‍

Efficiently processing datasets with loops

Loops are a convenient and efficient way to process structured data in your apps. 

With a single loop, you can easily run dynamic actions for hundreds of pieces of data in just one automation. 

If you’d like to learn more about automating your work with Zapier, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Like any modern web app, Zapier is constantly updating their product to make sure they keep up with the changing demands of their user base. 

‍

In this past year alone, we’ve seen multiple overhauls of Zapier’s visual layout. Even in the months since we posted our updated Beginner’s Guide in March of 2024, Zapier’s taken on a new coat of paint and rearranged some important items. 

‍

In this quick blog post, we want to quickly show you the three most important changes that you should know about as a no-code builder.

‍

We’re also going to explain why, in spite of these updates, you can still use our beginner’s guide and other tutorials from the same time period if you’re just picking up Zapier for the first time right now. 

Overall changes: an aesthetic overhaul

We’ll begin with a quick “before and after” of Zapier’s automation builder.

‍

Below, you can see the old Zapier UI from March 2024.

‍

‍

And in the next screenshot below, you can see today’s Zapier (October 2024). 

‍

‍

In general, you can probably see the biggest changes here are just some surface-level aesthetic choices. There’s a new color scheme with purple accents and several other minor visual tweaks. 

‍

However, these visual changes aren’t our main focus in this post.  

‍

Instead, let’s examine the three most important functional changes. 

‍

None of these updates amount to huge changes for Zapier users, but it’s definitely good to be aware of them if you’re just getting started and exploring tutorials from our blog – or looking at Zapier help docs that are about a year old. 

Connecting accounts in the “Setup” tab

Comparing the old UI to the new, there’s still a Zap canvas in the central part of the screen, with a simple summary of all your automated steps. 

‍

When you click on any of those steps, a menu pops open on the right where you can configure it. 

‍

Here’s where you’ll see the first key difference. 

‍

For a long time, Zapier divided each step into 4 tabs – “App & Event”, “Account”, “Trigger”, and “Test”. 

‍

‍

(And of course, that “Trigger” tab would be replaced with “Action” if the step was an action.)

‍

Now, there are just three tabs in the configuration menu: Setup, Configure, and Test. 

‍

‍

With this layout, you can choose your event and sign into an account all in the same “Setup” tab. 

‍

Despite this change, app authorization works the exact same way as before; it’s just in a slightly different place. 

Finding (and enabling) dynamic data

This next change is arguably the most important one. 

‍

With the new updates, inserting dynamic data into your Zaps works a little differently than it used to. 

‍

Previously, a window including all of the data retrieved by previous steps in your automation would automatically pop up whenever you selected an input field. 

‍

‍

Now, you’ll need to enter a quick command to access dynamic variables. 

‍

You can either type forward slash (/) directly into the field, or you can click on the plus button. 

‍

‍

Either one will pull up a list of your Zap’s available dynamic data. 

Configuring a field to accept dynamic data

If you don’t see the plus button in a field where you expected to use dynamic data, then it probably means that field is set to “static”. That means it’s configured to only accept inputs from a pre-set list of options. 

‍

‍

However, there’s an easy option to change “static” fields into “custom” fields that will accept variables.  

‍

Just click on the three dots button next to the field, and choose “custom”. 

‍

‍

Once you’ve set the field to ‘custom’, you can use either forward slash (/) or click on the plus to enter dynamic data. 

Revealing hidden tooltips

Finally, let’s take a look at one more key change to the Zap builder UI: the location of tooltips. 

‍

Previously, Zapier included useful help text under each field to let you know what data you needed to provide, or what each option in a list would do. 

‍

‍

This help text was often a little long and frequently needed to be truncated. In that case, you would have to click on “more” to see the full text.

‍

Now, with the new Zapier UI, there’s no help text immediately visible. However, it’s not gone. 

‍

‍

To see help text in the new Zapier UI, just click on the underlined title of any field.

‍

‍

Then, a tooltip will pop up in its own window.

‍

And just note that you do have to click on the titles to see the help text; it won’t appear on a mouseover.

Navigating new settings

Those are all the functional changes we wanted to go over in the main Zap builder. 

‍

But before wrapping up this blog, we also want to note that the settings menu on the left has been substantially overhauled recently. 

‍

‍

It won’t change how you build Zaps initially, but it may change how you configure them after they’ve been built. 

‍

If you’d like to learn more about using these updating settings, you can check out our earlier blog post on the topic. 

Keeping up with a changing platform

They say the only thing that’s constant in life is change – and that certainly applies to no-code automation, too. Zapier looks quite a bit different today than it did less than a year ago, and it’s a near certainty that this post will be outdated within a year, as well. 

But for now at least, Zapier still works in largely the same way. You’ll just need to keep an eye out for a couple of things that have moved around. 

If you’d like to explore additional news, tips, and tutorials about Zapier and other low-code platforms, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

Airtable, the popular low-code database and workflow platform, has just announced a game-changing update for enterprise teams and anyone working with large datasets. Their new "HyperDB" feature now allows users to access and work with databases containing tens of millions of records, far surpassing Airtable's previous record limits. 

‍

Record limitations in Airtable

‍

Airtable bases on even the highest Enterprise plans are limited to 500,000 records per base and 250,000 records per table. This has made working with datasets in the millions of records very challenging, often requiring complex workarounds.

‍

One of the main reasons users have choses tools like Google Sheets over Airtable has been the ability to work with larger datasets, as Sheets supports over 10 million cells. 

‍

How HyperDB Overcomes Record Limits

‍

HyperDB is opening up new possibilities for working with large datasets by allowing you to store your data in an external database like Snowflake or Databricks, which have much higher record capacities. You can then use HyperDB to sync specific subsets of that data into your Airtable bases and tables as needed.

‍

‍

For example, you could store a 50 million record product database in Snowflake, but sync only the most recent 100,000 records to an Airtable table. 

‍

Alternatively, you could sync records for different years or product categories into separate bases. The synced data remains connected, so you can easily change which records are pulled in.

‍

Use Cases for HyperDB

‍

While individual Airtable tables and bases still have the same record limits, HyperDB makes it much more straightforward to connect them to massive external datasets. This enables large-scale data processing, automation, and reporting use cases that were previously impractical in Airtable alone.

‍

For instance, you might use HyperDB to power operational analytics, or machine learning workflows that require analyzing millions of records.

‍

HyperDB is currently in beta and only available to customers on Airtable's Enterprise plans. Interested companies will need to contact Airtable to request access.

‍

Additional new features in Airtable

‍

In addition to HyperDB, Airtable has unveiled new features to help users manage and update their bases. 

‍

App Library: centrally controlled templates for your team

‍

The new App Library allows you to create reusable base templates that teams can copy for their own purposes while maintaining a dynamic link to a central reference version. 

‍

‍

For example, you could build an "OKR Tracker" app library for use across your company. Each department can spin up their own OKR tracking base from the template, customized to their needs. But the template owner can still push updates and gather roll-up reporting from all linked bases.

‍

This promotes consistency and data consolidation even as teams adapt bases to their unique workflows.

‍

App Sandbox: risk-free experimentation and seamless updates

‍

The App Sandbox provides an isolated environment to test changes to a base before deploying them to your team's live version. This allows you to experiment with new formulas, automations, and app structures without the risk of disrupting active workflows.

‍

‍

The sandbox is a game-changer for Airtable developers and builders, as it allows them to iterate more quickly and safely, whether improving their own team's bases or building client projects as an Airtable consultant or agency.

‍

Organization Branding: add your company’s logo and colors to Airtable

‍

Enterprise plan customers can now customize their bases with their own logo and colors in place of Airtable's default branding. This is especially useful for agencies, consultants, and others building client-facing applications on the Airtable platform. 

‍

‍

Branded bases provide a more polished, professional, and trust-inspiring experience when sharing them with external stakeholders, while still leveraging all of Airtable's flexible database and workflow functionality under the hood.

‍

The App Library and branding options are currently available for all Enterprise customers. The Sandbox is in beta, with Business and Enterprise plan users able to request access.

‍

Learn more about all recent Airtable updates here

Our take

‍

At XRay, we're extremely excited about HyperDB and the possibilities it opens up for our clients. Many of the companies we work with have huge proprietary datasets they want to build workflows around, and HyperDB is going to make that possible. HyperDB is a major step forward in making Airtable an ideal platform for enterprise-scale low-code application development. 

‍

We're also very pleased to see Airtable making investments in other areas that will help larger organizations adopt the platform, like the app sandbox and expanded branding options. We expect these updates to make a real difference for companies looking to run mission-critical processes on Airtable.

‍

Reach out to us if you'd like to discuss how your team can make the most of these new Airtable capabilities. As an Airtable Gold Partner, we're here to help you build powerful, scalable low-code workflows.

If you’d like to learn more about software like Airtable, Zapier, Softr, and more, check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

AI and no-code automation are a natural pairing. 

‍

You can take the ability of an AI like Claude to generate and analyze text, and combine it with automation’s time-saving potential to produce a fast, efficient, and effective workflow. 

‍

That’s why in this tutorial, we’re going to show you how to automate Claude by Anthropic with Zapier. 

‍

To begin, we’ll give you a brief overview of what Zapier is and what it does for the uninitiated.

‍

Then, we’ll show you how to set up your Anthropic account and connect it to Zapier. 

‍

Finally, we’ll walk you through the process of building an example automation that sends a prompt to Claude to extract action items from any email that you tag with a custom label. 

What is Zapier?

Before we dive into automating Claude, we’re going to quickly cover what Zapier is and how it works for anyone who might not be familiar with it. 

‍

Zapier is a no-code automation provider. It lets you build automated workflows by connecting your web apps together. 

‍

When an event happens in one app, one or more automatic events will be performed in other apps. 

‍

The automation we’ll build in this tutorial

For instance, in the automation – or “Zap” – we’re going to build today, everything kicks off when we apply a custom label to an email in Gmail. 

‍

‍

Then, Zapier sends a preconfigured prompt to Claude that includes text from the email’s body. 

‍

‍

Finally, the Zap automatically sends an email back to us with the action items that Claude extracted from the original email. 

‍

‍

In short, it will give you a summary with a single click. And you don’t need to write any code to build it. 

Connect thousands of apps to Claude with Zapier

You should know that even though we’re just gonna connect Gmail and Claude in this tutorial, Zapier supports over 7,500 apps right now - and that number’s always growing.

‍

The odds are good that Zapier can automate the apps you already use every day. 

‍

While Zapier does have a free plan, you’ll need a paid plan to build automations like this with more than two steps. 

‍

You can learn more about using Zapier in our Zapier beginner’s guide here. If you’re brand new to Zapier, we’d strongly recommend checking out that guide before proceeding with this tutorial. 

‍

Note that Zapier’s UI has already changed in the few months since we posted that guide. However, everything still works the same under the hood. 

Set up your Anthropic account for API access

In order to start automating Claude with Zapier, you’ll need to set up an Anthropic Console account, and you’ll need to set up an API billing plan. 

‍

That’s because Zapier is going to access Anthropic through its API, or “Application programming interface”. 

‍

You can learn more about what APIs are here. 

‍

Adding an API billing plan to your Anthropic account 

Go to console.anthropic.com and sign in with your existing Anthropic account, or create a new one. If you’re making a new account, you’ll need to provide a couple details about your organization.

‍

Once you’re signed in, click on “Settings”. Then choose “Plans & billing”. 

‍

New Anthropic console accounts always begin on the “Evaluation” plan, which doesn’t grant access to Anthropic’s API. 

‍

Click on “Select Plan” to see your options. 

‍

‍

You can either go with the “Build” plan, which will require you to load up your account with prepaid credits, or the “Scale” plan, which will bill you monthly for your activity. 

‍

‍

The “Scale” plan will require you to contact Anthropic’s sales team, so to get set up fast, it’s best to just go with the “Build” plan. 

‍

You’ll just have to provide some details about how you plan to use Claude, and where you’re located. Then, you can add a payment method and credits. 

‍

Note that you won’t need to pay any kind of subscription fee. You’ll only be charged for the credits you add to your account. 

‍

The good news is that running prompts through Claude’s API is generally going to be very cheap - often just a cent or less for each one. 

You can learn more about pricing details on this pricing page. 

‍

Once you’ve created a console account and added a payment method, you’ll be ready to connect Anthropic to Zapier. 

Connect your Anthropic account to Zapier

To connect Anthropic to Zapier, start by signing in to Zapier. 

‍

Click on “Apps” in the left hand menu. Then, choose “Add connection”. 

‍

‍

Search for “Claude” or “Anthropic” to find the app, and choose it from the list of results. 

‍

‍

You just need one piece of information to authorize Zapier to access your Anthropic account - an API key. 

‍

‍

To create an API key, go back to the Anthropic console. You can follow the handy link that Zapier provides. 

‍

Then, on the “API keys” page in the Settings tab, just click on “Create Key”. 

‍

‍

Give your key a unique and descriptive name so that you know what it’s for. You’ll be able to track usage by key, so we’d recommend naming it something like “Zapier Key” so you can see how much API activity is coming from your Zaps. 

‍

‍

Choose a workspace for your key, and click “Create key” to finish. 

‍

‍

Then, make sure to copy the key before closing this window, since Anthropic won’t reveal it to you in full again. 

‍

Paste it into Zapier, and click Continue.

‍

‍

Now, your Anthropic connection in Zapier is all set, and you’re ready to start building an automation. 

Creating a Zap to automate Anthropic (Claude)

To build a Zap that automates Claude, begin by clicking on “Zaps”. Then, create a new Zap. 

‍

As mentioned earlier, the automation that we’re going to build will start whenever we apply a custom label to an email in Gmail. 

‍

Even though we’re going to be building an automation that uses Gmail specifically, the same principles will apply to any other app that you want to connect with Claude. 

‍

So if you want to build an automation that launches prompts from Microsoft Teams, or from Hubspot, the information in this tutorial will still help you out. This is just one specific example that illustrates some general concepts. 

Preparing your test data

We’ll start by making our test data first. For our example, we’ll create a custom label in GMail – “Summarize with Claude” – and add it to an email so that we’ll have some test data to work with as we’re building the automation. 

‍

Always make sure you have some test data to use before you build your automation. 

‍

‍

Below, you can see a perfect candidate for this automation – exactly the kind of wordy, demanding email you’re probably all too familiar with. 

‍

‍

Create a trigger for your automation 

Once your test data is ready, add a trigger to your Zap. The “trigger” controls when the Zap runs. 

‍

Choose an app for your trigger. For our example, the triggering event will occur in GMail. 

‍

‍

Next, select a specific event that will trigger your automation. In our example, that event will be a newly labeled email. 

‍

‍

Note that you will need to connect your Google account to automate it with Zapier, but you won’t need to mess around with API keys or anything like you had to with Claude. 

‍

You’ll just need to sign in to Google and click a button to approve Zapier. 

Configuring your trigger

Next, configure the specific settings for your trigger. Configuring the GMail trigger in our example is extremely simple. 

‍

All you have to do is identify the label that you want the automation to trigger for. In this case, that will be our custom label - “Summarize with Claude”. 

‍

‍

With the trigger configured, you can give it a test to pull in some data. This test data is required to build the other steps in the automation, and will show you what kind of information Zapier can retrieve and work with. 

‍

‍

After a successful test, you can see a long list of data, like the email addresses of the sender and recipient, the email body, and many other items that you typically won’t need to use. 

Send an automated prompt to Claude

With a successfully tested trigger, you can now add a step to your Zap that sends a prompt to Claude. 

‍

Add an action, and choose “Anthropic (Claude)” as the app. 

‍

‍

For the time being, there’s only one choice for an action: “Send message.” Pick that, and choose your connected Anthropic account if it’s not already selected. 

‍

‍

Then, click “Continue.”

‍

Unlike the trigger, there are several options to configure here - and you can find even more by setting “Show Advanced Options” to “True.” 

‍

Configuring your automated prompt

As you configure your prompt, note that only fields marked with an asterisk are required. You could just fill out your message and leave all of the other settings to their defaults if you want to get set up quickly. 

‍

However,  configuring your action a little more precisely usually pays off. 

User message

First, you’ll see the “User Message” field. Here, you’re just going to enter a prompt pretty much like you would if you were chatting with Claude directly. 

The most important difference is that you can also include dynamic data retrieved from previous steps in the Zap.

‍

For example, we can include the body of the email that triggered the Zap so that Claude can analyze it and find action items. 

‍

‍

System

Under “System,” you can provide instructions for the role that Claude should assume. Entering additional information here can help to shape Claude’s style and approach to answering your user message.

Model

Next, you can choose a model to use. 

‍

‍

You can check out this page for more information about pricing and performance to inform your choice. 

‍

We’d recommend going with a faster model for any Zap that uses Claude. Zapier won’t wait indefinitely for a response, so you’ll want to prioritize getting a reply quickly, even if that means you can’t always use the latest and greatest model.

Memory key

In the “Memory Key” field, you can provide a unique key to connect your conversation with previous ones. 

‍

For example, if we use a key like “EmailActionItems”, then Claude will use previous conversations with the “EmailActionItems” key as context, until it hits its context limit. 

This can be a good way to give your conversations more depth and to improve the results over time.

Image

Next, there’s an option to upload images so Claude can use “vision” to analyze them. 

‍

We won’t attach an image to this prompt in our example, but keep in mind that you could add images here as dynamic data retrieved from earlier steps - such as an image attached to the original email. 

Max Tokens

In the “Max Tokens” field, you can set a token limit if you'd like to keep the response short and your API charges low.

‍

Temperature, Top K, Top P

Finally, there are settings for “Temperature,” “Top K,” and “Top P.” 

All of these settings are about changing the predictability or creativity of the AI. 

‍

We won’t get into all of the technical nuts and bolts about how these work – if you’re interested in the details, you can explore this IBM documentation to learn more about how these factors influence an AI’s output. 

‍

However, for most users, reading the descriptions in Zapier will be fine. 

‍

In our example, we’ll set “Temperature” to .7 to make the answers a little more creative, but not all the way at the max of 1. 

‍

Testing the prompt

Once you’ve configured all of your settings, you can give your prompt action a test. Note that running a test on a Claude prompt in Zapier will actually send the prompt to Anthropic’s API, and will incur a small charge on your account. 

‍

‍

Our test was a success, and we can scroll down past the user message to see Claude’s answer.

‍

Not bad. This definitely produces a much more readable list of tasks than the original email. 

‍

But you can always tweak your prompt, system message, and settings to fine tune your results if you’re not satisfied.

Send Claude’s response to another app

Now that we’ve got a response from the AI, the last thing we need to do is make sure the output actually gets sent to us somewhere we can easily read it. 

‍

Again, you could send it to thousands of different apps, but we’ll just go with the logical choice and send it right back to our Gmail inbox.

Add another action to your Zap

Add another action to the Zap, and choose Gmail as the app again. For the action, we’re going to choose “Reply to email.” 

‍

‍

Then continue to the configuration. 

‍

First, you need to identify a thread to reply to. 

‍

Change the field to “Custom” instead of “Static,” and insert the thread ID retrieved in the trigger. 

‍

‍

For the “To” email, just enter your own email.

‍

‍

By only entering your own email here, you can make sure that the reply doesn’t get sent to the original sender. 

‍

CC and BCC are optional, but you will need to enter a “from email” - that can be your email address again. 

‍

You’ll also need a “from name.”

‍

‍

Keep the body type as plain, type up a brief message, and insert Claude’s response from the previous step. It’s the “Response content text” variable. 

‍

‍

You don’t need to add any labels or attachments, so you can continue and test. 

‍

‍

After your successful test in Zapier, check your Gmail inbox to see the final message. 

‍

‍

Now, you can turn the Zap on and start using it. 

Once it’s on, it will trigger after you add a label to an email. However, it won’t be immediate. 

Zapier will check for new emails at set intervals, which you can see here in the left-hand menu of the Zap builder. 

‍

‍

Depending on your plan, you may have to wait up to 15 minutes. But on higher subscription levels, you can reduce the wait down to just a minute. 

Combine Anthropic and Zapier for automated AI prompts

Anthropic and Zapier are a winning combination that will help you line up your daily tasks and knock them down in record time. Getting started is easy, and after you’ve completed this tutorial, you can try connecting Anthropic with any other apps that you want. 

To learn more about optimizing your workflows with AI and automation, check out the other tutorials on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.

If you’ve opened Drive recently, you’ve probably seen an announcement about new shortcuts.

‍

‍

Google Drive has always supported keyboard shortcuts for common actions like creating new documents and folders, but these new shortcuts work a bit differently. 

‍

We’ve put together this quick tutorial to help Google Drive power users get a handle on these new keyboard commands. 

Google Drive Keyboard Shortcuts - Creating new items

To create any new item, such as a doc, a slide deck, or a folder, begin by pressing Control + C on Mac. 

‍

If you’re using a Windows PC, you’ll use Alt + C instead. 

‍

‍

Then, press a key that corresponds with whatever you want to create, like “F” for a new folder. 

‍

‍

NOTE - you will need to release Ctrl + C or Alt + C before pressing the single letter that will create a new item. 

‍

Once you’ve created a new folder, you can use the Enter or Return key to open it up. 

‍

Creating new docs, slide decks, sheets, and forms

The Control + C or Alt + C combination is how you’ll start creating new items. The following combinations will use the same pattern to let you easily make presentations, forms, spreadsheets, and docs. 

‍

Mac OS

• New Google Slides Presentation = Control + C, then P

• New Google Form = Control + C, then O

• New Google Sheet = Control + C, then S

• New Google Doc  = Control + C, then T

‍

Windows

• New Google Slides Presentation = Alt + C, then P

• New Google Form = Alt + C, then O

• New Google Sheet = Alt + C, then S

• New Google Doc  = Alt + C, then T

‍

If you’re not sure what the keys are, you could just use the POST acronym to remember them:

‍

Presentation

Form

Sheet

DocumenT

‍

If you ever need a refresher, or want to explore some of the other shortcuts available in Drive, you can find the full list of all Google Drive shortcuts here. 

Creating items in Google Drive with URL shortcuts

Alternatively, you can also create new items in Google Drive by simply typing a “.new” address into your browser’s URL bar. 

‍

For instance, navigating to the “docs.new” URL creates a new Doc in your Drive. 

‍

‍

Similarly, “sheets.new” creates a new Google Sheet. 

‍

‍

Note that you will need to be signed in to Google Drive for these URL shortcuts to work. 

‍

Whether you use these shortcuts or the new hotkeys, using these little tricks can make your daily tasks in Google Drive a little bit faster. 

Level up your productivity with workflow automation 

If you’d like to take your productivity to the next level, check out XRay’s workflow automation tutorials on our blog or on our YouTube channel. 

‍

Each week, we’ll show you new ways to automate the software you use every day, like Google Drive, Slack, ChatGPT and more. 

‍

You can also learn more about the services we offer for small businesses and enterprises, as well as the free products we’ve built to help our fellow workflow designers. 

And if you want to get the latest updates about our content, you can also follow XRay on Twitter, Facebook, or LinkedIn.

‍

Sometimes, the data that Make retrieves from your automated apps doesn't have all the information you need. 

‍

When you need to add some numbers, text, dates, or other info to your Make scenarios, you can use Variables to fill in the gaps. 

‍

In this tutorial, we’re going to show you how to create and use scenario variables in Make, formerly Integromat.  

How to create a scenario variable in Make

First, let’s go over the process of creating scenario variables in Make. 

‍

Scenario variables are custom, user-created variables that only exist within the scenario you create them in. You can set their names and values to whatever you’d like. 

‍

To create a custom variable, add a “Tools” module to your scenario. 

‍

‍

Then, choose “Set variable.”

‍

‍

You can enter anything you’d like for your variable name. You can even have spaces if you want, but we’d recommend making them all one word, and just use CamelCase like in the screenshot below for clarity.

‍

‍

You’ll have to use your variable name to retrieve it later, so it’s best to make sure you can easily and accurately type it out each time. 

‍

Then, provide a value for your variable. 

‍

How to set a Make variable to a numeric value

NOTE: Directly entering a number as the variable value will not make it a valid number. 

‍

If you type text directly into the “value” field, the text will be a string. This means that if you type “52”, for instance, your variable’s value will be a piece of text that says “52”, not the number 52. 

‍

If you try to add 1 to the variable in a subsequent module, the value will become a string that says “521”, not 53.

‍

If you want to make your variable a number, then use the parseNumber function, and enter your number as the first argument (before the semicolon). 

‍

‍

If your number has a decimal, enter the decimal separator as the second argument (after the semicolon). Otherwise, just leave it blank. 

‍

You’ll need to use a similar formula in order to enter a valid date. You would use the parseDate function, enter your date, and so on. 

Referencing your variable in a subsequent module

Once you’ve configured your variable’s title and value, click “OK” to commit your settings. 

‍

Whenever this module runs, it will set the value of your variable to whatever you specified. 

‍

Then, you can use the variable in subsequent steps just like any other data processed in Make. For instance, we’ll use Slack like a makeshift dev console in the screenshot below, and send the variable as part of a simple message. 

‍

‍

Scenario variables are a great way to create data that isn’t present in the automations you’re building. 

‍

‍

They’re especially useful if you don’t want to store that information in an external source like Airtable or SmartSuite. 

Update a variable’s value

To automatically change the value of your variable while your scenario is running, just add another “set variable” step. Specify the variable name, using the exact name of the variable you want to update. 

‍

Remember, it’s case sensitive. 

‍

Then, enter the new value you want to assign. You can enter an arbitrary value directly, or you can calculate the new value based on the earlier value. 

‍

For instance, we’ll just add “1” to the previous value of the variable. 

‍

‍

Now, whenever this module runs, it will add “1” to the value of the variable. 

‍

We’ll add another Slack step and run the scenario to see the full output. 

‍

‍

The first message shows the original value of the variable – 52 – and the second message shows its updated value of 53. 

‍

Retrieve a variable’s value from another path

Let’s wrap up this tutorial on scenario variables with a quick look at how you can retrieve the current value of any variable. 

‍

In many cases, you can often just grab the variable from the last module that set its value, but this won’t work if the variable was last set in a different path. 

‍

For instance, in the scenario pictured below, the Slack module in the second path can’t directly retrieve the variable set in the first path. 

‍

‍

To retrieve the value of a variable that’s been set in a different path you can use “Get variable” to retrieve the current value of the variable. 

‍

‍

Just enter the name of the variable to find it. As always, you’ll need to write it exactly the same way as the module that sets its value. 

‍

‍

Then, you can use the variable retrieved by the “Get” step in any subsequent modules in the path. 

‍

‍

Just remember to consider the order that your paths will run in. They’re numbered (1st, 2nd, etc) so you can tell which paths Make will attempt to run first. 

‍

‍

Also, consider which paths you expect to run for each piece of data your scenario processes. Depending on the timing, you could always add a short delay step, to ensure the proper order of things. 

Augment your Make scenarios with custom variables

Scenario variables are a useful little tool that can make it much easier to build the automations you want, even if you’re not getting all the data you need directly from your apps. 

If you’d like to learn more about automating your work with Make, Zapier, and other no-code tools, be sure to check out the other posts on our blog or our YouTube channel. You can also follow XRay on Twitter, Facebook, or LinkedIn.