Track Leads and Customers from AI Tools in Pipedrive
AI assistants increasingly send people to business websites, yet a visit alone does not show whether that visit produced a qualified lead, a paying customer, or revenue. This guide explains how to attach source data to every form submission so Pipedrive records show which AI tool and landing page were involved.
Why AI visits need to reach the CRM record
Website analytics can count visits arriving from tools such as ChatGPT or Perplexity, but those counts stop before the commercial outcome. They do not show how many visitors submitted a form, became customers, or generated revenue.
The practical fix is to preserve the referral information at the moment of conversion. When a visitor submits a form, the source details travel with the name, email, and other submitted fields into Pipedrive. Once those details sit on the Person or Deal record, reporting can connect AI activity to pipeline and sales results instead of treating it as anonymous website traffic.
Four-step setup checklist
Install the attribution snippet on the website, then add four hidden fields to each lead-capture form:
- Channel
- Channel Drilldown 1
- Channel Drilldown 2
- Channel Drilldown 3
Most form builders support this through a hidden field type or by hiding a text field in its settings. The same fields should be used consistently across forms so records can be compared in Pipedrive.
When a visitor arrives, the attribution code examines the available referral signals and assigns a channel such as AI Referral, Organic Search, Paid Search, or Paid Social. That result is stored in the browser. On form submission, the code writes the channel details into the hidden fields.
The form tool then captures those values with the submission. Supported form, scheduling, and chat products named in the captured evidence include Gravity Forms, Contact Form 7, Wix Forms, Webflow Forms, Typeform, Jotform, and Zoho Forms. From there, a native form-builder integration or a tool such as Zapier can create a Person or Deal in Pipedrive.
What the captured fields mean
Each lead carries two kinds of information: how the visitor reached the site and which page they landed on.
For an AI referral, the channel fields can be read as follows:
- Channel = AI Referral
- Channel Drilldown 1 = the AI tool, such as ChatGPT
- Channel Drilldown 2 = the referring domain, such as chatgpt.com
- Channel Drilldown 3 = None, because no deeper detail is available for that referral type
Landing-page fields add the entry page and its group. In the captured example, the landing page is smithandco.com/blog/tax-planning-tips and the landing page group is /blog. That distinction matters because it shows which site content AI tools are reading, citing, and sending visitors to.
Four reports to build after capture
Once the fields are present on Person and Deal records, four reports give a clearer picture than visit counts alone:
- Leads by Channel: monthly lead volume split across AI Referral, Organic Search, Paid Search, Paid Social, and other channels. This shows AI's share of pipeline and whether that share is growing.
- Customers by AI Tool: new customers split by the tool that sent them, allowing ChatGPT, Claude, and Gemini to be compared on paying outcomes rather than mentions or visits.
- Close Rate by Channel: the proportion of leads that become customers within each channel. Volume can mislead; a smaller AI lead stream may close at a much higher rate than paid or organic traffic. The captured illustration contrasts 35% for AI Referral with 10% for Organic Search and 7% for Paid Search.
- Revenue from AI by Landing Page: customers and revenue attributed to each page reached through AI referrals. This identifies the posts and pages that lead to buyers and gives a practical basis for deciding what to create next.
These reports can be built in Pipedrive Insights or in a connected analytics tool such as Looker Studio, Databox, or Microsoft Power BI.
Final check before you trust the numbers
Before relying on the reports, check that every form uses the same four hidden fields, that test submissions populate them correctly, and that the form-to-Pipedrive mapping creates the intended Person or Deal record. Then review the four reports together: lead share shows scale, customers by tool shows commercial value, close rate shows lead quality, and revenue by landing page shows which content deserves more investment.
Frequently asked questions
Why is website analytics not enough to measure AI referrals in Pipedrive?
Website analytics can show visits from AI tools, but not whether those visitors became leads, customers, or revenue in Pipedrive. Capturing channel and landing-page data on the form submission and passing it to the Person or Deal record closes that gap.
Which hidden fields should a form include?
Add hidden fields named Channel, Channel Drilldown 1, Channel Drilldown 2, and Channel Drilldown 3 to each lead-capture form. The attribution code then fills those fields when a visitor submits.
Why does Channel Drilldown 3 show None for AI referrals?
Channel Drilldown 3 is None for an AI referral because there is no deeper level of detail to capture at that level.
Where can the captured AI data be reported?
Use Pipedrive Insights, or connect a tool such as Looker Studio, Databox, or Microsoft Power BI to report on the captured fields.