ChatGPT says it handles a fifth of Google's search volume. Plan for both
OpenAI says ChatGPT search volume is about 20% of Google's. Treat the claim with care, but build AI search visibility alongside classic search now, not after the data is clean.
What happened
On 3 November 2026, 9to5Google reported that OpenAI says ChatGPT's search volume has reached about a fifth of Google Search's. The remark came from Adam Fry, OpenAI's ChatGPT consumer product lead, at an OpenAI press event. He described it as "roughly one-fifth (20%) of the volume of queries of Google." 9to5Google credits The Verge with reporting the statement.
The article is careful about what the number means. OpenAI did not say which Google volume it used for the comparison. It defines a search as any prompt that triggers a web search, which is not the same unit as a Google query. 9to5Google also notes that Bing holds a single-digit share of search, and that ChatGPT entered AI-powered search in October 2024, about two years ago.
Two other points in the report matter to marketers. First, Google has acknowledged "lower monetization" from its AI search push. Second, Fry said most people do not seem to mind ads in ChatGPT, and that OpenAI could show more of them. He also said users value ChatGPT's personalisation and its understanding of them.
We have not independently verified OpenAI's figure, and the article says that OpenAI gave no detail on the Google volume it compared. We treat it as a claim by an interested party.
Why it matters
The number is soft, but the direction is hard to ignore. A company with an interest in the answer said ChatGPT's search is a fifth the size of Google's, measured in its own way. We would not plan around the number. We would plan around the fact that a share of your buyers now start research in an assistant, and that OpenAI thinks the share is large enough to announce.
Two channels now need two kinds of measurement. Classic search gives you rankings, clicks and conversions. An assistant gives you an answer that may or may not name you, and may or may not link to you. Without measurement of the second, you are managing half of the picture and cannot tell whether work on it pays.
Paid options are arriving in the same place. Fry's comments that ads are tolerated and could increase suggest that the assistant's answer may become paid space as well as earned space. Your organic position there is worth building before the paid one becomes the default way in.
Monetisation pressure will shape the experience. Google has acknowledged lower monetisation from its AI search push. When a business under that pressure changes how answers and ads appear, B2B teams should expect more change, not less, and should avoid tying their plan to one layout.
Our view is that the cheapest response is also the most robust one: clear, specific, well-sourced content on your own site, accurate profiles elsewhere, and a regular check of what assistants say about you.
What we would do
- Write your buyer questions down. List the thirty to fifty questions your buyers ask in the weeks before they contact a supplier, grouped by topic. This becomes both your content plan and your test set.
- Test them in the assistants your buyers use. Run the set monthly in ChatGPT, Google's AI features and any others relevant to your market. Record whether you appear, how you are described and which sources are cited.
- Fix what they get wrong. If an assistant misstates your services, find the sources it relies on and correct them, starting with your own site.
- Publish answers, not slogans. Create pages that answer a buyer's question directly, with evidence, named authors and dates. Keep them current.
- Add referral traffic to your reporting. Track visits from assistants separately from search and social, and record the change over several months before you judge it.
- Hold your classic search investment steady. Do not move budget on one company's claim. Decide the split when you have your own numbers.
The aim is to know where you stand in both channels, so that the next claim, from whichever company, is a prompt to check your own data.
Questions we are asked
Should we take OpenAI's 20% figure at face value?
No. The figure comes from an OpenAI press event, OpenAI did not say which Google volume it compared against, and it counts any prompt that triggers a web search. That is not the same as a Google query. It is still a useful signal that a meaningful number of buyers research through ChatGPT, which is reason enough to measure your own presence there.
Does this mean we should reduce our investment in Google search?
Not on this evidence. The claim compares volumes on different definitions, and Google remains the larger channel. The sensible response is to add measurement and content work for AI assistants without cutting what already produces pipeline. Review the split once you have your own data on referral traffic and on how often assistants mention your brand.
How do we measure visibility in ChatGPT?
Start with a fixed set of questions your buyers ask, across the topics where you want to be found. Run them regularly in the assistants your buyers use, record whether and how your brand appears and which sources are cited, and track changes over time. Pair this with referral traffic in your analytics. Treat results as directional, because answers vary between runs.