Million Dollar YearGet the free app

How to capture leads from a custom GPT into your CRM

By Max Warnault · 8 August 2025 · 5 min read

In short

You can make a free custom GPT collect leads. The GPT asks for a first name and email. A Relevance AI tool takes those two inputs and creates a contact in your CRM, like GoHighLevel or HubSpot, with a tag that names the GPT. Relevance AI gives you an API key and an OpenAPI schema, and you paste both into a new action in the GPT. The tag can then trigger any automation you want.

A free custom GPT makes a good lead magnet. People use it, they get value, and then they leave. Unless you capture their details, you have nothing to follow up with.

I wanted my brand builder GPT to ask each user for a first name and email at the end, then add them to my CRM and my community. I found an easy way to do that with ChatGPT and a tool called Relevance AI. Here is the whole setup, step by step.

What you need to turn a custom GPT into a lead magnet

  • A custom GPT in ChatGPT that people use for free (mine helps build a brand).
  • A Relevance AI account.
  • A CRM. I use GoHighLevel. If you use another CRM, such as HubSpot, the setup works the same way.

The flow is simple. The GPT asks for the first name and email. It passes them to a tool in Relevance AI. The tool creates the contact in your CRM with the name, the email and a tag.

Step 1: Create the tool in Relevance AI

Once your Relevance AI account is set up, there is a section for custom GPT actions. That is where the GPT connection happens. But you have to build a tool first, because the custom GPT action uses that tool.

Set the inputs

  1. Go to Tools and create a new tool. Mine is called brand builder GPT.
  2. Set the inputs. These are what your agent passes into the tool. For this, the inputs are the first name and the email address.

I plan to build one tool that works for all my GPTs. That way I don't have to create a new one every time I make a GPT.

Add the step that creates the contact in your CRM

  1. Add a step that sends the person to your CRM. Mine adds them to GoHighLevel.
  2. If you use HubSpot, search the steps for the CRM options and pick the one that creates or updates a contact in HubSpot. There are other steps there too, like assigning a contact, creating a task or updating a company. You only need the contact step.
  3. Connect your CRM account. For GoHighLevel, that means adding your location ID.
  4. Map the information you want in the CRM: the name, the email and a tag.

The tag matters. Mine names the GPT the lead came from. I would also like the tool to pick up the GPT's name on its own, but I'm not sure that's possible yet. For now the tag is set by hand.

  1. Save the changes.

Step 2: Get the API key and the OpenAPI schema

Go back to the custom GPT actions section in Relevance AI and scroll down. Your new tool will be listed there.

From there you can generate two things:

  • An API key.
  • An OpenAPI schema to paste into your GPT.

Step 3: Add the action to your custom GPT

In ChatGPT, open your GPT or create a new one. For this walkthrough I made a test GPT.

  1. In the instructions, tell the GPT what to do. For example: ask the user for their first name and email to get access to free AI resources.
  2. At the very bottom of the configuration, click to create a new action.
  3. For authentication, go back to Relevance AI, copy the API key, select API key in ChatGPT, paste it and save.
  4. Copy the schema from Relevance AI and paste it into the action. The tool now shows up in the GPT.
  5. Add your privacy policy.

Step 4: Test the GPT and check your CRM

Now test the whole thing. When I said hello to my test GPT, it asked for my first name and email to give me free access to AI resources. I typed my name and a test email.

Next, ChatGPT asked whether I allowed it to talk to Relevance AI. I agreed. The GPT then confirmed I was all set and that my free AI resources were activated.

Then I went to the contacts in my CRM and refreshed the page. The test email address was there. At the bottom of the contact was the tag "GPT brand builder."

Use the tag to trigger your follow-up automations

The tag is what makes this a lead generation system, and not just a list of emails. Each GPT gets its own tag, so you always know where a lead came from. Your CRM can use that tag to trigger any automation you want: add the person to your community, send an email sequence, or anything else you already run.

This is the kind of system we build at Million Dollar Year. A piece of code runs while you sleep and grows your list without adding hours to your day. You build the GPT once, connect it once, and every new user who shares their details lands in your CRM with the right tag.

Key takeaways

  • Add a closing instruction to your free custom GPT that asks for the user's first name and email in exchange for something, like free AI resources.
  • Build a Relevance AI tool with two inputs (first name and email) and a step that creates the contact in GoHighLevel, HubSpot or your own CRM.
  • Tag every contact with the name of the GPT it came from, so you always know the source.
  • Paste the Relevance AI API key and OpenAPI schema into a new action in your GPT, and add your privacy policy.
  • Test the GPT with your own email, then check that the contact and tag show up in your CRM.
  • Use the tag to trigger your follow-up automations, such as adding the lead to your community.

Questions and answers

How do I collect emails from a custom GPT?

Tell the GPT in its instructions to ask for the user's first name and email. Then add an action that sends those details to a tool in Relevance AI. That tool creates the contact in your CRM.

Does this work with HubSpot or only GoHighLevel?

I use GoHighLevel, but you can use another CRM. In Relevance AI, add the step that creates or updates a contact in HubSpot instead of the GoHighLevel step, connect your account and map the name and email.

What do I paste into the custom GPT action?

Two things from the custom GPT actions section of Relevance AI. Paste the API key in the action's authentication, then paste the OpenAPI schema. You also need to add a privacy policy.

Do I need a separate Relevance AI tool for every GPT?

Right now I have one tool per GPT, with a tag that names the GPT. I plan to build one tool I can reuse for all my GPTs, so I don't have to make a new one each time.

What happens after the lead is in my CRM?

The contact gets a tag, such as "GPT brand builder." You can use that tag to trigger any automation you want in your CRM, like adding the person to your community or starting a follow-up sequence.

For 6-figure coaches and agencies

Build a million dollar business, solo, with systems and AI.

Start with the free app: your year planned down to today, your numbers in one place, and exactly what your goal takes.

Get the free app

Free · No card · On the web and on your iPhone

Keep reading