OpenAI says ChatGPT users dismissed ads at half the rate in mid-2026 compared to when the company launched its advertising business in February. The claim comes from Chief Revenue Officer Denise Dresser, speaking at Cannes Lions, and it has not been confirmed by any third-party measurement firm.
Ad dismissal rate is OpenAI’s chosen proxy for relevance. The company treats fewer dismissals as evidence that users find the ads useful and contextually aligned with what they are doing in the conversation. That is a reasonable proxy, but it is OpenAI’s own interpretation of OpenAI’s own data. No independent source has measured engagement, click rates, conversion, or brand recall across ChatGPT’s ad surface.
The absence of external measurement matters precisely because the 50 percent figure is the only number on the table. When Google reports AI Overviews impression data, analysts can cross-reference it against Similarweb traffic shifts, publisher revenue logs, and Search Console click curves. ChatGPT’s ad inventory has no equivalent external audit trail yet. The figure should be read as an investor-ready signal, not a verified benchmark.
What the development does confirm is structural. ChatGPT’s ad surface is maturing faster than many search and PPC teams have noticed. The February launch was quiet enough that most agencies treated it as an experiment to monitor later. Dresser’s Cannes statement is an attempt to reframe the timeline: the experiment is producing data, and the company wants brands to treat it as a live channel.
For search and PPC teams, the comparison to the traditional SERP ad surface is instructive. A Google Search ad sits alongside organic results after a keyword fires. A ChatGPT ad sits inside a task-completion session, appearing when a user is actively solving a problem rather than browsing query results. That intent context is closer to what Google’s Performance Max calls a high-intent moment, except the session is conversational and the user has explicitly asked for help. Lower dismissal rates, if accurate, would suggest users in that context tolerate relevant ads better than they tolerate sidebar banners or post-click redirects.
Dresser described the format as being about usefulness, a framing that distinguishes it from interruptive display inventory. Whether that holds as OpenAI scales the program and financial pressure to fill inventory increases is the question that current data cannot answer.
OpenAI is simultaneously competing for enterprise AI contracts against Anthropic and others, which creates pressure to show that consumer ad revenue can diversify the business. The 50 percent figure, timed to a major advertising festival, is partly a signal to the brand industry that ChatGPT is worth a budget line.
Search Engine Land reported Dresser’s remarks on June 24.
Search and PPC teams should treat ChatGPT’s ad inventory as a surface to brief into planning cycles now, before holding data from the platform becomes table stakes for client conversations.
Reporting by Anu Adegbola, Search Engine Land, published June 24, 2026.
