DalanovumAI visibility can be measured. Most people do it wrong.
Three studies from October cover more than 200 websites and twelve months of data. They agree: AI is a brand channel, and it should be measured like one.
By Fawzi Aiboud Nygren, Dalanovum · Published
Short answerAI assistants send about 2.5% of the traffic that organic search sends, and no study has shown that visibility in AI answers leads to more branded searches. Measure AI as a brand channel: the same buyer questions every month, saved answers, and ask buyers how they found you.
In October 2026, three studies came out in the same week.123 Search Engine Journal brought them together on 6 November.4 They agree. AI assistants send little traffic, and most AI visibility metrics have never been tested against what buyers actually do.
Visits from organic search and AI assistants, 95 websites
Visitors from AI assistants buy more often. In Marie Haynes' data, ChatGPT visitors converted at 13.8%, against 3.02% from Google.2 But the visits are so few that Google still drove 96.5% of conversions.
Share of visits that converted
For B2B products, AI traffic converts worse
SEO Works split the sites by customer type.1 Companies selling B2B products got the largest share of AI traffic, but it converted worse than search traffic on six of ten sites. For B2B services, AI and search were close.
Sites where AI visitors converted worse than search visitors
No one has found the hidden effect
The usual argument is that buyers see your name in an AI answer and search for it later. Kevin Indig looked for exactly that in twelve sites' data.3 Direct traffic had grown 140.9%, but engaged direct traffic had fallen 1%. The growth was bots. On one site, 198 of 409,000 direct visits were engaged.
He then ran 632 tests of AI visibility against direct visits and branded searches for four brands. When he shifted the data to the wrong dates, several links were as strong as with the right ones. His conclusion: an effect could exist, but he could not isolate it.
Measure AI like a billboard
A billboard is not measured in clicks. It is measured in how many see it and what they remember. AMEC, the international association for the measurement of communication, published data quality principles in July 2026 that fit AI visibility well.5 This is how we measure:
- Ask the same buyer questions every month and save the answers. A screenshot of an AI answer is an anecdote.
- Keep mentions, citations and recommendations apart. They are not the same thing.
- Test the number against branded searches and deals, and against wrong dates as a control.
- Ask buyers how they found you, in forms and at the first meeting.
- Put the money where the evidence meets. Sites that grew in AI traffic also grew in search traffic.1
That is how our industry panel works: the same buyer questions every night, with the answers saved, so changes show over time.
Frequently asked questions
How much traffic do AI assistants send?
About 2.5% of the traffic organic search sends, according to SEO Works' data from 95 websites. Marie Haynes found 0.73% for all AI assistants together.
Does AI traffic convert better?
Often, but not for everyone. For B2B products, AI traffic converted worse than search traffic on six of ten sites.
Does AI visibility lead to more branded searches?
No one has shown it yet. Kevin Indig found no stable link in 632 tests.
So is AI visibility unimportant?
No. Buyers use AI when they compare vendors. But it should be measured as a brand channel, not a traffic channel.
Sources 5
- SEO Works, AI referral traffic benchmark. Data September 2025 to August 2026, 95 GA4 properties. ↑
- Marie Haynes, quoted by Search Engine Journal: GEO is a distraction, 21 October 2026. 100 sites, 12 months. ↑
- Kevin Indig, Growth Memo: Does AI drive more direct traffic?, 19 October 2026. ↑
- Search Engine Journal, AI visibility isn't a sham, most ways of measuring it are, 6 November 2026. Where we found the studies, the same day. ↑
- AMEC, Data Quality Principles, July 2026. ↑