OpenAI said on October 5 that it’ll begin testing a visual ad format inside ChatGPT later this month, placing clearly labeled display units alongside image-generation results for U.S. users on the Free and Go plans. The format is launching with a limited initial group of advertisers, per TechCrunch, and arrives bolted to something more consequential than any creative unit: a measurement stack.
That second half is the real announcement. More than a dozen attribution and incrementality partners are shipping support simultaneously. Click attribution runs through AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge, and Tenjin. Fospha, Measured, and INCRMNTAL are named for full-funnel work. Hightouch, Tealium, and LiveRamp cover conversion-data pipes. Haus, Measured, and WorkMagic are running early geo-based incrementality experiments. DoubleVerify and Integral Ad Science are piloting brand-suitability evaluations without access to real user conversations. OpenAI says the ads will be clearly labeled, kept separate from the image being created, and won’t influence ChatGPT’s answers.
The context explains the kitchen-sink partner list. Since ChatGPT Ads launched in February, advertisers have complained to Digiday about three things: poor measurement, few formats, and thin inventory. This release addresses two of them in one swing.
The third, inventory, remains the structural problem. Zeno Group’s evp paid social Shamsul Chowdhury told Digiday that ChatGPT runs roughly 2.5 billion prompts a day against Google’s 15 billion searches, with only one ad placement on ChatGPT versus multiple on Google, which keeps supply tight. OpenAI’s own framing cites 1.2 billion weekly users and positions ChatGPT as “the largest AI-native platform.” Reach is enormous; slots are scarce. Scarcity sets price, and price sets who gets in.
Which is where the access pattern starts to matter. The recent addition of CRM-list targeting carried a 25,000-user floor that walls off most five-to-thirty-person brands, and before that sponsored agents turned ads into conversational sales surfaces behind their own thresholds. The visual format isn’t self-serve yet. Smaller teams that want in when it broadens should pre-wire a compatible attribution partner now, because the measurement side is finally real.
Yahoo Finance flags the open questions worth watching: ad frequency per generation session, whether targeting eventually draws on conversation context, and whether OpenAI publishes independent brand-suitability audit results before expanding beyond the U.S. The staging looks less like a product launch than a dress rehearsal for an auction. First the measurement rails, then the formats, then the inventory expansion that lets the auction clear at the price OpenAI actually wants.
Sources
- https://openai.com/index/new-chatgpt-ads-format-and-measurement/
- https://techcrunch.com/2026/10/05/openai-launches-visual-ads-that-appear-alongside-image-generation-results/
- https://digiday.com/marketing/openai-moves-to-make-chatgpt-ads-more-measurable-and-more-visual/
- https://www.bleepingcomputer.com/news/artificial-intelligence/openai-will-show-visual-ads-in-chatgpt-while-you-generate-images/
- https://finance.yahoo.com/media-advertising/articles/openai-putting-visual-ads-next-184547191.html