Track AI referral leads and customers from Typeform forms
AI assistants such as ChatGPT, Claude and Gemini increasingly send visitors to business websites. Standard analytics can show that traffic arrived, and AI visibility tools can show mentions or citations, but neither tells you how many of those visitors became leads, how many became customers, or what they were worth. This guide shows a practical way to carry AI referral data through a Typeform submission so you can answer those questions in your own reporting.
What you are actually measuring
There are three different questions people often blend together:
- Did an AI tool mention or cite us? This is a visibility question, answered by AI tracking tools.
- Did an AI tool send us a website visitor? This is a traffic question, answered by web analytics.
- Did an AI-referred visitor become a lead, then a customer, and how much revenue did they produce? This is the question this guide addresses.
Only the third question connects the source of the visit to a business outcome. To answer it, the referral information has to travel with the form submission into whatever system holds your lead and customer records.
The approach described here uses Attribution Analytics to detect and categorise the visitor's source, write that data into hidden fields on your Typeform, and pass it onward to your CRM or a spreadsheet. Reports then group leads and customers by channel and by the specific AI tool.
Setting up the capture: four steps
1. Install the tracking snippet and add hidden fields
After signing up for Attribution Analytics, you receive a code snippet to place on your website. Installation instructions are linked from the product help centre.
Next, add these hidden fields to each Typeform you use for lead capture:
- Channel
- Channel Drilldown 1
- Channel Drilldown 2
- Channel Drilldown 3
- Landing Page
- Landing Page Group
In Typeform, hidden fields live under the Logic tab in the 'Personalise with data' option. You then adjust the embed code for your form — the exact change differs between the standard, full-page and popup embed types — so the tracking data can be passed in.
2. Let the tool classify the visitor
When someone arrives on your site, Attribution Analytics examines a range of signals to work out where they came from, in the same general way analytics platforms do. It assigns the visit to a channel such as AI Referral, Organic Search, Paid Search or Paid Social, and stores that classification in the visitor's browser.
3. Capture the data with each response
If that visitor later completes a Typeform — a quote request or demo booking, for example — the stored referral details are submitted along with their name, email and other answers. You can view the values beside each response in the Responses area of your Typeform account, and include them in lead notification emails. Sending them to a CRM keeps the source attached to the lead record.
4. Report on the results
Once the data reaches your CRM or spreadsheet, you can group and chart it using the CRM's built-in reporting or your spreadsheet's charting tools. More capable reporting platforms such as Tableau, Looker Studio or Microsoft Power BI can also be used. If you connect a CRM to an AI assistant, you can ask a plain-language question such as how many AI Referral leads arrived this month.
What each field will contain
The table below maps the six hidden fields to their meaning, using a worked example. Imagine an IT services provider that supports small offices. A prospect asks ChatGPT for a managed IT provider, clicks through to the provider's site, and submits a quote request form.
Four reports worth building
The value of the captured data depends on how you slice it. These four views cover the most common questions.
Leads by channel. Group monthly lead counts by channel — AI Referral, Organic Search, Paid Search, Paid Social, Direct and so on. Use it to size AI against your established channels and decide whether it warrants more of your attention.
Customers by AI tool. Group monthly new-customer counts by the referring AI tool. This lets you compare, for example, ChatGPT against Claude and Gemini. Tools that produce actual customers, rather than only mentions or visits, indicate where your audience is and where visibility work is most worthwhile.
Close rate by channel. Divide customers by leads for each channel. AI-referred visitors often arrive with their questions already answered, so their close rate can be higher than other channels. Treat any specific figure as something to measure in your own data rather than an expectation to assume.
Revenue from AI by landing page. Attribute customers and revenue to the page each AI-referred visitor first landed on. Pages that consistently produce buyers are a reasonable signal for what content to create next.
A short checklist before you rely on the numbers
- Confirm the hidden fields exist on every lead-capture form, not just one.
- Check that the embed code change matches the embed type you actually use.
- Submit a test response and verify the values appear in the Responses view.
- Confirm the CRM mapping puts each field in a reportable property, not a free-text note.
- Decide how you will treat leads with no detectable source before you start comparing channels.
- Agree a reporting period and keep it consistent so month-to-month comparisons mean something.
Where to go next
Capturing AI referral data in Typeform turns an open question — is AI actually producing business for us? — into a reportable one. The setup is short, and the reporting options range from a spreadsheet to a full BI tool depending on what you already use.
Trial length, plan pricing and setup-time claims are commercial terms that must be confirmed by the site owner before publication.
| Field | Example value | What it tells you |
|---|---|---|
| Channel | AI Referral | The broad source category |
| Channel Drilldown 1 | ChatGPT | Which AI tool sent the visitor |
| Channel Drilldown 2 | chatgpt.com | The domain the visitor arrived from |
| Channel Drilldown 3 | None | No deeper level of detail exists for an AI referral |
| Landing Page | clearpathit.com/services/managed-it-small-business | The first page the visitor saw |
| Landing Page Group | /services | The section of the site that page belongs to |
Values shown are illustrative; the domain and page will reflect the actual referring tool and your own site.
Frequently asked questions
Does this tell me how often AI tools mention my brand?
No. Mention and citation tracking is a separate measurement. This setup records visitors who arrive from an AI tool and then submit a form, so it measures leads and customers rather than visibility.
What happens if a visitor's source can't be determined?
The classification depends on signals available in the browser. Visitors whose source cannot be identified will not be labelled as AI Referral, so treat the AI figure as a floor rather than a complete count.
Do I need a CRM to do this?
No. A spreadsheet can hold the submitted values and provide charting. A CRM makes it easier to connect leads to closed customers and revenue.
Why does Channel Drilldown 3 show None?
An AI referral has no deeper level of detail beneath the tool and its domain, so the third drilldown is empty for these visits.