Most AI search advice is written by people who have never measured AI traffic on a site they own. This post is the opposite: raw GA4 referral data from the 12 sites in my own portfolio that receive the most AI traffic, pulled on 26 August 2026, covering the 90 days from 28 May to 26 August 2026. This is a refresh of the original 2 August pull, same 12 sites, same method, updated numbers below.
The headline: 4,717 sessions arrived from AI assistants across these 12 sites in 90 days, and 90% of them came from ChatGPT.
Method, before anything else
Numbers without a method are marketing. Here is mine:
- Source: Google Analytics 4, one property per site, queried via the GA4 Data API on 26 August 2026.
- Window: 28 May to 26 August 2026 (90 days).
- Definition: a session whose
sessionSourcematches chatgpt, perplexity, claude, gemini, copilot or openai. This captures click-throughs from AI assistants, not impressions inside AI answers. The smallopenaisource is counted in the ChatGPT column, since it is the same ecosystem. - Selection: the same 12 sites tracked in the original 2 August study, chosen for the highest AI referral traffic from my wider portfolio of 70+ live sites.
- Sites are labelled by sector rather than domain, because several earn affiliate revenue and I would rather publish the data than invite cloners.
I audited the popular third-party statistics in this space separately in my AI search statistics fact-check. Most failed a primary-source check. Everything below is my own data.
The results
| Site (by sector) | AI sessions (90d) | ChatGPT | Copilot | Perplexity | Claude | Gemini |
|---|---|---|---|---|---|---|
| Health and wellness | 1,357 | 1,145 | 71 | 79 | 62 | 0 |
| Education data | 993 | 930 | 40 | 20 | 3 | 0 |
| EV charging | 573 | 520 | 43 | 5 | 2 | 3 |
| Document templates | 595 | 589 | 3 | 1 | 0 | 2 |
| Property data | 259 | 231 | 14 | 10 | 4 | 0 |
| Remote work and travel | 155 | 143 | 0 | 10 | 2 | 0 |
| Water quality | 221 | 176 | 34 | 9 | 2 | 0 |
| Mobile coverage | 176 | 143 | 30 | 3 | 0 | 0 |
| Garden equipment reviews | 260 | 251 | 9 | 0 | 0 | 0 |
| QR tools | 80 | 77 | 3 | 0 | 0 | 0 |
| This site (SEO consulting) | 48 | 45 | 0 | 1 | 1 | 1 |
| Fitness | 0 | 0 | 0 | 0 | 0 | 0 |
Totals: 4,717 sessions. ChatGPT 4,250 (90.1%), Copilot 247 (5.2%), Perplexity 138 (2.9%), Claude 76 (1.6%), Gemini 6 (0.1%).
Finding 1: ChatGPT is 90% of the AI referral market
Every site with meaningful AI traffic has the same shape: ChatGPT first by a wide margin, everything else sharing scraps. If your AI visibility strategy is spread evenly across five assistants, you are misallocating 90% of the effort. Optimising for how ChatGPT selects and cites sources is, today, most of the job.
Finding 2: AI referral traffic survived a Google core update that crashed organic clicks
This is the finding the first pull could not show, because it needed a second measurement window to appear. The health and wellness site and the education data site both lost most of their Google organic clicks to a core update during this period. One fell to near-zero daily clicks. Their AI referral sessions did not collapse with them. Google re-ranks a page against the current SERP. AI systems cite from a source list built up over time, and losing the first does not automatically remove a site from the second. Full breakdown on the AI Visibility Results page.
Finding 3: Gemini referral traffic barely exists
Six sessions. Across twelve sites. In ninety days. Google's assistant either answers without citing, cites without generating clicks, or both. Do not let anyone sell you a "Gemini optimisation" line item on the strength of referral traffic, because there is barely any to optimise for. Google's AI Overviews affect your Google organic clicks instead, which is a different problem with different measurement.
Finding 4: condition-led data sites win, commercial sites lose
The top of the table is entirely sites that answer specific factual questions: health conditions, school catchment data, charger locations, water hardness by area. The review site in the table earns a fraction of what the data sites earn. AI assistants cite sources that resolve questions with checkable facts. If your site's core content is "buy this", you will need a data layer to earn citations.
Finding 5: Copilot behaves differently, and referral counts undersell it
Copilot sends modest referral traffic here, but it concentrates on utility sites (EV charging, water quality, mobile coverage: the three highest Copilot counts outside health). Copilot also surfaces brands inside answers where no click ever happens. One documented case from this portfolio: a healthcare company found and hired this practice through a Bing Copilot citation, a lead that referral analytics attributed to nothing. Referral data is the floor of AI visibility, not the ceiling.
Finding 6: the consultant's site nearly doubled, the fitness site dropped to zero
This site earned 48 AI referral sessions in this window, up from 28 in the first pull. That is still a fraction of what the data-led niche sites earn, but it is moving in the right direction as more primary data gets published here. The fitness site went the other way: 13 sessions down to 0. Both changes are small enough to be noise on their own, and both are reported here rather than only the number that flatters the argument. Publishing primary data nobody else has remains the single lever that moves this number, up or down.
What this means if you want AI traffic
- Weight your effort toward ChatGPT citation patterns and treat the rest as secondary until the data changes.
- Publish checkable primary data in your niche. It is the single strongest citation magnet in this dataset.
- Do not assume a Google ranking drop means an AI visibility drop. Measure the two separately, they do not move together.
- Measure the floor and the ceiling separately. GA4 referral data for clicks, plus brand-presence checks inside AI answers for the invisible layer. My free AI visibility check covers the second part.
- Ignore vendors quoting industry statistics they cannot source. Ask for their own data. If they have none, you have learned what you needed to know.
If you want this measurement done properly for your business, with a competitor benchmark and a prioritised plan, that is exactly what the AI Visibility Audit is. If you are still working out what this discipline involves, start with what a GEO agency actually does.
Related Articles
- AI Visibility Results - the current snapshot of this same data, refreshed on its own page
- AI Search Statistics: Fact-Checked - the popular numbers traced to their primary sources
- What Is LLM Optimisation? - the practice behind these referral numbers
- Optimise Content for AI Search - content patterns that earn citations
- Why Your Brand Is Not Appearing in ChatGPT - diagnosis for invisible brands
