Transform data with Airtable AI

Learn how to use Airtable AI to automate data generation and transformation tasks.

Tags Streamline Icon: https://streamlinehq.com
Wrench Streamline Icon: https://streamlinehq.com
Uses
People Man Graduate Streamline Icon: https://streamlinehq.com
Intermediate

Managing data cleanliness and accuracy can be a manual, time-consuming task for any team, and if you have your data stored in a local Excel file, this can be even more challenging.

This is where a tool like Airtable can help. With its new AI features, it can automate your data management tasks across a slew of use cases. If you’re not familiar with Airtable, it’s a cloud-hosted relational database tool that has many AI capabilities built directly into the data and automation layers of the product.

In this tutorial, we’re going to walk through how you can use all of the AI features in Airtable to automate data generation and transformation tasks, freeing you up to focus on higher-leverage work.

You’ll need:

Steps:

  1. Add templated AI fields to your data table
  2. Add AI fields from scratch to your data table
  3. Use ChatGPT to write formulas
  4. Use AI in Airtable automations
Products Give Gift 1 Streamline Icon: https://streamlinehq.com

Try before you bite?
This one's on us.

Sign up for a free account to view our free courses

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Add templated AI fields to your data table

To start, make sure you have an Airtable Base created with at least one data table and a few records added to the table. For this example, we’re going to start with a table of meeting records.

To add an AI field, use the Plus button by the column header and select the AI field type at the bottom of the pop-up.

Select the Generate Text option, name your field, and then select the “Start with a template” option below. Since we have meeting data, we’re going to create a field called Meeting Summary.

You’ll notice there is a large list of options that Airtable has already pre-developed to make using AI within your data table quick and easy. We’re going to select the Meeting Summary option from the available templates and define a few parameters.

We’ll set the starting point to be our Meeting Notes field and we’ll also include the Meeting Name field to provide extra context. After these two parameters are set, we’ll click the Generate Preview button to see a sample of our output.

If you’re happy with the output, click the Use Template button to create the field, or you can provide additional custom instructions by adding examples or providing additional prompts before re-generating the preview output.

Once you’ve clicked the Use Template button, make sure to click the gear icon on the ensuing pop-up. There you can enable the option to generate AI outputs automatically, toggle on/off the AI model optimization setting, as well as set a numerical level of randomness.

💡 Tip: We typically leave the bottom two options set to the defaults. For “Generate automatically,” you should decide whether or not you want the AI field to run every time new data is added to the fields linked to the AI generation field, or if you want to trigger updates manually.

You can check out more of the available AI-templated fields by going through the same process above in new columns in your data table. There are many templated options to choose from, including categorization, translation, sentiment analysis, and more.

Add AI fields from scratch to your data table

Instead of using Airtable’s templated AI fields, you can write custom AI prompts from scratch, referencing fields from your Airtable table within the prompt. To do this, add a new AI field to your Airtable data table and select the “Start from scratch” option.

Once you click that option, a free text field will appear where you can write your prompt. In this example, I’m prompting the AI to categorize my meeting record based on my previously AI-generated meeting summary field. Make sure to review the Preview of the output below your prompt to ensure it matches your desired result before creating the field.

Sample prompt:

Categorize the below meeting into one of the categories provided. Only output the category and nothing else.

Categories: Design, Strategy, Marketing, Product

Meeting: {Meeting Summary}
💡 Tip: If you are prompting the AI with a categorization or labeling prompt, it’s often helpful to include a sentence like “Only output the [output desired] and nothing else.” Otherwise, you might get additional commentary from the AI that is undesired.

Similar to the templated AI field, you can click the gear icon on the field pop-up to set the various parameters like autogeneration, model, and randomness.

Use ChatGPT to write formulas

Alongside Airtable’s AI fields, you can use ChatGPT as a co-pilot to write highly complex formulas to copy/paste into the Airtable formula field.

Why might you want to do this? Using Airtable’s AI fields can be more expensive than traditional formula fields that come at no additional cost. If you have a data transformation requirement that can be accomplished with a traditional formula, we recommend going that route first before using an Airtable AI field. This isn’t always possible, which is why Airtable’s AI fields are so helpful, but it’s a recommended option when you have more straightforward data transformation needs.

In this example, we’ve added a Meeting Date field to our meeting records. I want to transform this date field from its current format of MM/DD/YYYY to a day-of-the-week format (e.g. Tuesday). This can be done using an Airtable formula field.

Let’s hop over to ChatGPT and prompt it to write an Airtable formula field for this day-of-the-week output.

Sample Prompt:

I have an Airtable field "Meeting Date" with date data. I want to create a formula field "Day of the Week" which transforms that date into the day of the week format, like "Tuesday." Can you provide me the formula for this?
💡 Tip: If you reference the exact name of your Airtable fields in your prompts, ChatGPT will write the exact formula you need. You can then copy/paste it into Airtable and it will work with no adjustments needed. Otherwise, you’ll need to update the field reference in Airtable.

Click the Copy Code button in the formula output in ChatGPT, go back to Airtable, and drop the formula into a new formula field input box.

You can use this workflow for writing many formulas in Airtable to transform your data without having to dig into the Airtable formula documentation.

💡 Tip: Sometimes ChatGPT will invent a function that does not exist. When that happens, let it know and it’ll generate another version for you to test. If you can’t get a formula to work, that’s a good moment to hop over into an Airtable AI field to perform the data transformation.

Use AI in Airtable automations

Lastly, we can use AI in Airtable automations. This is helpful if you want to transform data coming from a third-party application before adding it to your data table or have AI generate data for certain fields based on an action or time-based trigger.

In this example, we’re going to have our AI step assign a meeting record to a user based on the meeting type, and we’re going to have this happen every time a meeting record is created.

To start, navigate to the Automations section of your Airtable Base and select the “When a record is created” trigger. On the right window, select your Table.

Below the trigger, click the “Add advanced logic or action” button and select the “Generate with AI” option.

💡 Tip: There are no templated AI prompts in the Automations section of Airtable. So for this action, you will always need to write your prompts from scratch.

Enter your prompt for the AI generation step in the Prompt field on the right-side window. In this example, we’re having AI assign a Meeting record to a User based on the Meeting Type of the record.

Sample prompt:

You assign Meetings to Users based on Meeting Type. Only output the User name and nothing else.

Meeting Type: {Meeting TypeValue}

Matching Rules:

IF Strategy = Mark

IF Design = Lina

IF Product = Dmitri

Lastly, we want to add a final step to our automation to update a record with our AI-generated output. To do this, add an “Update record” step below your “Generate with AI” step. In the right side window, select your table, select the Airtable record ID from the first step, and add a new field with the AI-generated response as it’s input from the second step.

In this example, we’ve selected a User field to hold the Response from the AI-generated step.

From there, you can generate a preview of the output and enable your automation for your Airtable Base.

Nice work! You now know all of the ins and outs of Airtable’s AI features to start generating and transforming data directly in your Airtable data tables.

This tutorial was created by Garrett.

Get full access

✔️ All 100+ courses & tutorials in our catalog
✔️ New content added weekly
✔️ Private community access
✔️ No subscription, $150 paid once
✔️ Expense it using this template. Or get a team account.
✔️ 30-day refund policy. No questions asked
Join 3,107 learners from companies like Microsoft, Coca Cola, NBA, Adobe & Google

More tutorials like this

View all

If you scrolled this far, you must be a little interested...

Start learning ->

Join 3,107 professionals already learning