Turning AI-Search Traffic Into Leads: A Guide to Website Visitor Identification

Website visitor identification is the practice of resolving anonymous site traffic into named companies and people, so a session count becomes a lead list. If you're doing AEO — getting cited in ChatGPT and Perplexity, ranking in Google AI Overviews — you're winning the hard half of the funnel. You're getting found. The question nobody's analytics answers is the next one: who actually showed up, and did you capture them?

For most businesses the honest answer is "no idea." Google Analytics says 412 sessions and a bounce rate. It can't tell you that a dental group read your pricing page twice this week, or that a returning visitor from a law firm just hit /contact and left. You paid to earn that traffic. Then it walked out the door anonymous. This post is about closing that gap — what visitor identification is, how the tools actually differ, and how to turn AI-search traffic into booked leads.

AEO is really good at the top of the funnel. Get cited, get found, pull qualified people to your site. But "found" is not "captured." There's a loop most teams never close:

Found → the AI answer names you, someone clicks through. Identified → you know which company and person that was. Captured → you get their details, on your terms. Booked → they're on the calendar.

AEO tools and web analytics handle the first step and stop. The middle two — identified and captured — are where the money leaks. You did the expensive work of earning attention, and then the visitor stays a number in a chart. Fixing that is a different job than getting found, and it needs a different tool.

There are two different things people mean by "identifying a visitor," and they matter for how much you can trust the data.

### Company-level identification

When someone visits from a corporate network, their IP address can often be matched to the organization that owns it. That's how you learn "Beaverbrook Builders was on your solutions page," even if you never learn the individual's name. This is reverse-IP resolution, and it's deterministic — it's a lookup, not a guess.

### Person-level identification

Resolving the actual human — name, email — is a higher bar. The clean way to do it is on signals the visitor gave you: they filled a form, or they clicked a tracked share link that carries an identity. Some tools try to resolve individuals from third-party identity graphs (cookie pools, data brokers). That can work, but it's probabilistic, patchier, and raises privacy questions depending on where your visitors are.

The distinction to hold onto: deterministic identification is a fact you can act on; probabilistic identification is a hint. When you're about to spend a salesperson's time on a lead, you want facts.

There are four buckets of tools people reach for here, and each does a slice of the job. Here's the honest version of who does what.

### Web analytics (Google Analytics, Plausible)

Great at aggregate traffic — sessions, sources, bounce, conversions you've defined. But it's anonymous by design, it doesn't tell you which company or person a session was, and the data lives on someone else's cloud. It measures traffic; it doesn't identify or capture anyone.

### De-anonymization tools (RB2B, Leadfeeder / Dealfront, Koala)

This is the category built for exactly this problem, and it's genuinely useful. Leadfeeder (now Dealfront) pioneered company-level reveal from reverse IP. RB2B leans into US person-level reveal. Koala blends intent scoring for sales teams. If all you need is "which companies hit my site," these do it well. Their limits: most stop at identification (they hand you a list, not a captured lead), they don't map each visitor's full journey through your site, they run on their cloud with your visitor data in it, and person-level reveal can get probabilistic. Pricing is usually several hundred dollars a month on top of everything else.

### Conversational / chat tools (Intercom, Qualified, Drift)

Strong at capture — they'll book a meeting off a live chat — and increasingly AI-first. But they identify a visitor mostly once that visitor engages the widget, they're expensive, and again it's their cloud, their model.

### All-in-one, self-hosted (Cite-Met Pulse)

This is the bucket we build in, so read it with that in mind. The idea is to do the whole loop in one system on infrastructure you own: resolve the company and person, map the full per-visitor journey page by page, and capture the lead with built-in forms and a booker — deterministically, with the data in your own database instead of a vendor's cloud. The trade-off is that it's delivered as a managed service rather than a self-serve signup.

Regardless of which tool you pick, these are the questions that separate a real lead engine from a fancier analytics dashboard:

Does it show the person, not just the company? A company name is a start; a named contact with their journey is a lead.

Does it map the whole journey? Knowing someone visited is weak. Knowing they entered on your AEO landing page, read pricing twice, and came back three times tells you they're in-market.

Does it capture, or just report? A list you have to go chase is worth far less than a form-fill and a booked call tied to that visitor automatically.

Who owns the data? If your visitor data and leads live in a vendor's cloud, they're a subscription away from disappearing — and they may be feeding someone's model. Self-hosted, own-your-data is a real differentiator, especially for GDPR.

Deterministic or probabilistic? Ask directly. You want to know when an identification is a fact versus a guess.

Cite-Met Pulse is our answer to the whole loop in one place. It resolves corporate visitors to their organization automatically and resolves people the moment they fill a form or open a tracked share link — deterministically, never from a black-box model. It maps every visitor's journey entry-to-outcome, runs a live feed of who just got identified and how, and captures leads with forms and a meeting booker built into the same one-line embed. All of it self-hosted, so your leads live in your database and are never sold or used to train a model.

The deliberate choice: Pulse isn't a standalone signup. It's included for clients on Space & Story's $700+/mo managed AEO retainer — the same engagement that gets you cited in the first place also identifies and captures the traffic it earns. One team, one system, found through booked. If you want the standalone-tool route instead, the de-anon tools above are the place to start.

Whatever you deploy, judge it on lead outcomes, not vanity identification counts. Track a short list monthly:

Identified accounts (companies resolved from your traffic) Named contacts captured (form-fills + booked calls tied to a real person) In-market signals (returning visitors, pricing-page repeat views) Content that converts (which entry pages actually pull identified leads toward a booking)

If identifications climb but captured contacts stay flat, your problem isn't visibility — it's capture, and you should fix the offer and forms, not buy more traffic. If both climb, you've closed the loop: the AI-search work is earning traffic, and you can finally see and book the leads inside it.

Getting found in AI search is genuinely half the job. The other half is being able to answer "who showed up, and did we capture them?" — and that's a different tool than the one that got you cited. Company-level de-anon tools are a fine start. The bigger win is doing the whole loop — identify the person, map the journey, capture the lead — on data you own. That's the difference between "we got you traffic" and "here are the leads."