Skip to main content

AI Visibility

How to track traffic from ChatGPT and AI search

The Orchestra Team · October 14, 2026 · 5 min read

Is ChatGPT sending traffic to your website? You can increasingly answer that question — but you need to distinguish between two things that are easy to confuse: AI referral traffic and AI visibility. They aren't the same.

Tracking ChatGPT referral traffic

OpenAI's publisher FAQ states that publishers allowing OAI-SearchBot access can track ChatGPT referral traffic using analytics platforms such as Google Analytics, because referral URLs from ChatGPT search automatically include:

utm_source=chatgpt.com

That gives marketers a concrete way of identifying traffic originating from ChatGPT search.

What should you track?

At minimum:

  • Users — how many visitors arrive?
  • Landing pages — which pages receive AI referrals?
  • Engagement — do users interact with the page?
  • Conversions — do they sign up, request information, purchase, book or start a trial?
  • Revenue — for measurable transactions, what commercial value follows?

Compare AI traffic with traditional search

Don't just count visits — compare quality. An illustrative example:

ChannelSessionsConversion rate
Organic Search10,0002.1%
ChatGPT5004.8%
Paid Search3,0003.3%

A smaller channel can still matter if users arrive with strong intent. But don't assume AI traffic is high quality either. Measure it.

Why referral traffic doesn't measure AI visibility

Consider an answer that reads: "you could consider Company A, Company B or Orchestra."

The user sees Orchestra. That's a brand impression. But perhaps they never click. Your analytics reports zero sessions — while AI discovery clearly occurred.

That's the central limitation of referral-based measurement.

AI search creates a zero-click measurement problem

This isn't entirely new; search marketers have dealt with zero-click behavior for years. But generated answers intensify the challenge, because much more information can be delivered before a website visit becomes necessary.

So businesses need two layers of measurement.

Layer 1 — on-site: AI referrals, user behavior, leads, purchases, bookings, revenue.

Layer 2 — off-site AI visibility: brand mentions, citations, prompt coverage, competitor appearances, topic visibility.

Together they create a much fuller picture. How to measure AI visibility covers the second layer in detail.

Google is increasing AI visibility reporting too

Google launched dedicated generative AI performance reporting in Search Console in June 2026, initially for a subset of sites. The reports show generative AI impressions, pages, countries, devices and performance over time.

Note what that list doesn't include: clicks. The reporting is deliberately an exposure view — which is exactly why it's useful. It shows visibility that occurs without a traditional visit.

Bing has gone further into citation measurement

Bing Webmaster Tools' AI Performance reporting includes total citations, cited pages, grounding queries and citation trends.

So the measurement ecosystem is moving quickly from "how many AI visits did we receive?" toward "how does our information participate in AI discovery?"

Build an AI acquisition dashboard

Eventually your acquisition view might include Google organic, paid search, social, email, ChatGPT, other AI referrals, an AI visibility score, AI citations, and AI share versus competitors. That gives management a much more complete picture than sessions alone.

Don't optimize for clicks alone

There's a strategic consequence here. If AI systems mention your brand without sending traffic, that visibility can still contribute to brand awareness, later branded searches, direct traffic, buying consideration and offline enquiries.

Attribution becomes less linear. The objective should be visibility + engagement + conversion, rather than clicks alone.

Start measuring now

Even if AI referrals currently represent a small percentage of your traffic, creating a baseline now is valuable. You'll know which AI sources grow, which pages attract referrals, whether conversion quality differs, and when visibility changes.

When AI discovery becomes a major acquisition channel, you won't be starting from zero.

Referral analytics shows who clicked

Orchestra GEO reviews sampled brand, competitor and citation evidence across ChatGPT, Perplexity and Gemini before the click.

See how Orchestra GEO works

More to come

We’re publishing a new article on AI visibility and GEO every week. Leave your email and we’ll tell you when the next one lands.

New articles only. Unsubscribe in one click, any time.