Back to blog
Data & ResearchMar 20266 min read

How AI Identifies Your Leads: The De-Anonymisation Window

E
Elena TorresPublished Mar 2026

AI-referred visitors arrive pre-educated. They've already asked an AI assistant about your product, read the answer, and clicked through. They're not browsing. They're evaluating. They're the highest-intent visitors your site receives — and without the right tracking, they're completely anonymous.

The Attribution Gap Nobody Is Talking About

Every major analytics platform now reports a surge in 'Direct' traffic. Teams look at their dashboards, see direct traffic climbing while organic is flat, and chalk it up to increased brand awareness. What they're not seeing: a significant portion of that 'Direct' traffic is actually AI-referred.

When a user clicks a link from ChatGPT or asks Perplexity a question and follows a citation, the HTTP referrer is typically either empty or set to the AI platform's domain. Most analytics setups either drop this as Direct traffic or attribute it to a single 'AI' bucket with no further detail.

The De-Anonymisation Process

Fixing the attribution gap requires two layers working together:

Layer 1 — Source Identification

A correctly configured tracking script captures AI-referred sessions from all major platforms — ChatGPT, Perplexity, Google AI, Gemini, Claude, Copilot, DeepSeek, Grok — and tags each session with the referring AI platform, the landing page, and session data.

Layer 2 — Visitor De-Anonymisation

IP-to-company resolution converts an anonymous IP address into a named organisation. Combined with email lookup APIs, this surfaces the company name, the contact name, the email address, the industry, and the company's geographic data — all without the visitor filling out a single form.

Why AI-Referred Visitors Are Your Best Leads

Multiple independent studies have found AI-referred traffic converts at 2-3x the rate of organic search traffic. The reason is intuitive: organic search visitors are at the top of the funnel, discovering options. AI-referred visitors have already been through the discovery phase — the AI answered their category question, and they clicked through because your brand was in the answer.

They arrive knowing your name, having read a summary of what you do, and looking to validate whether you're the right choice. That's a fundamentally different intent level than someone who found you by searching 'project management software'.

Session Deduplication and Lead Scoring

Raw de-anonymisation produces noise — bots, one-second bounces, irrelevant visits. A properly configured lead attribution system applies deduplication (typically 1-hour windows to merge sessions from the same IP) and intent scoring based on session depth, pages viewed, and visit frequency.

A company that visited your site once for 12 seconds is not the same lead as a company that visited four times across ten days, viewed your pricing page twice, and arrived from Perplexity each time. Lead scoring surfaces the latter at the top of your pipeline and lets your sales team prioritise correctly.

What to Do With This Data

The three most effective actions once you have identified AI leads:

Hema's Inbound Leads feature includes the full tracking stack: AI referral attribution, IP-to-company de-anonymisation via Hunter.io, lead scoring, session timeline, and instant Slack + email alerts when a high-intent company visits. Available on the Business plan and above.

Related articles

Every new article, direct to your inbox.

No noise, just signal.