Track AI-sourced leads and customers in Webflow
Mentions in AI answers do not by themselves show business value. The practical question is whether those visits become enquiries, customers, and revenue. This guide explains how to attach referral information to Webflow form submissions so that information can be carried into a CRM and reported by channel, AI tool, and landing page.
Why capture the referral at submission time
A visitor who arrives after researching with an AI assistant often has a clearer brief than a cold visitor. That makes the source of a form submission commercially useful: it can change how the lead is prioritised, which tool is credited, and which pages deserve more content investment. Capturing the referral at the moment of submission is more reliable than trying to reconstruct it later from memory or from a separate analytics view.
Set up the form fields
First, add the attribution snippet to the website using the current installation instructions. Second, add six hidden fields to each Webflow form: Channel, Channel Drilldown 1, Channel Drilldown 2, Channel Drilldown 3, Landing Page, and Landing Page Group. Because Webflow's form builder does not offer a native hidden-field control, place an Embed block in the form, ideally just above the submit button, and paste in the supplied field code. Publish the site, then submit a test form and check that the values are present.
How the data travels with the lead
On each visit, the attribution tool examines available signals to classify how the visitor arrived, for example as AI Referral, Organic Search, Paid Search, or Paid Social, and stores that classification in the browser. When the visitor later completes a form, the stored values are written into the hidden fields. Webflow then saves those values alongside the name, email, and other details the visitor entered. They appear with the submission in the Forms section of site settings, can be downloaded as CSV, and are included in the 'New form submission' email. An integration platform such as Zapier or Make can forward them to a CRM or another system.
What an AI referral submission looks like
Two kinds of information are passed through: how the visitor reached the site, and which page they first landed on. Consider a property developer who asks an AI assistant for help finding a consultant for a contaminated land assessment, clicks through, and requests a quote. The submission would show Channel as AI Referral, Channel Drilldown 1 as the AI tool that sent the visitor, Channel Drilldown 2 as that tool's domain, and Channel Drilldown 3 as None, because AI referrals have no further drilldown level in this model. The landing page would be the service page the visitor first opened, grouped under its parent section. That combination shows both the referring tool and the content that attracted the visit.
Four reports to run first
Once the fields reach a CRM, most built-in reporting is sufficient for the core questions. Leads by channel compares AI Referral with Organic Search, Paid Search, Paid Social, and other channels month by month. Customers by AI tool splits new customers across tools such as ChatGPT, Claude, and Gemini, so effort can follow the tool the target market actually uses. Close rate by channel shows what share of leads from each channel become customers; in the source's illustrative example, AI Referral leads converted at a higher rate than Organic Search and Paid Search leads, consistent with visitors who have already done substantial research. Revenue from AI by landing page then shows which individual pages generate customers and income, pointing to the topics worth expanding. For more flexible views, the same data can be connected to a reporting tool such as Tableau, Google Data Studio, or Microsoft Power BI, or queried directly by asking a connected AI assistant how many leads came from AI referral in a given month.
A practical rollout checklist
Start with one high-value form, verify the six fields on a test submission, and confirm the values arrive in the CRM before rolling the setup out to every form. Then build the four reports above and review them monthly: volume tells you whether AI matters, tool splits tell you where to focus, close rates tell you what each lead is worth, and landing-page revenue tells you what content to create next.
| Field | What it records | Example for an AI referral |
|---|---|---|
| Channel | Broad source category | AI Referral |
| Channel Drilldown 1 | Specific AI tool that referred the visitor | ChatGPT (or whichever AI tool sent them) |
| Channel Drilldown 2 | Domain the visitor came from | chatgpt.com |
| Channel Drilldown 3 | Further detail where available | None |
| Landing Page | First page the visitor landed on | harlowenviro.com/services/contaminated-land-assessment |
| Landing Page Group | Parent section of the landing page | /services |
Field names and example values are taken from the captured source example; the None value for Channel Drilldown 3 reflects that AI referrals have no further drilldown level in this model.
Frequently asked questions
How do I add AI referral tracking to a Webflow form?
Add the attribution snippet to the site, then place six hidden fields in each Webflow form through an Embed block above the submit button. Publish the site and test a submission to confirm the fields arrive with the lead details.
Which fields are captured with each submission?
The captured set is Channel, Channel Drilldown 1, Channel Drilldown 2, Channel Drilldown 3, Landing Page, and Landing Page Group. Together they identify the broad channel, the specific AI tool and referring domain, and the first page visited.
Where can I see the AI referral data after a form is submitted?
Check the Forms section in Webflow site settings, where submissions can be viewed and downloaded as CSV, and check the 'New form submission' notification email, which includes the hidden-field values. From there the data can be forwarded to a CRM or other tools with an integration platform.
Which reports show whether AI referrals are worthwhile?
Useful starting reports are leads by channel, customers split by AI tool, close rate by channel, and revenue from AI referrals grouped by landing page. These show volume, quality, tool performance, and which pages generate value.