For the advertisers who have tried ChatGPT’s ad slots, the problem has not been reach. It has been proof. A marketer can place a message inside the product that 1.2 billion people use every week and still struggle to say what the spend did. On Monday, OpenAI moved to close that gap, adding visual ad formats and a set of measurement tools it says will let buyers see whether their money worked.
The change is the company’s clearest effort yet to turn advertising into a real business. ChatGPT has carried text-based ads in limited markets for months, rolling them out first in Europe and Southeast Asia. The new formats give advertisers room for images and richer creative, and they arrive alongside a measurement pixel that tracks what a user does after seeing or clicking an ad, down to a completed purchase on the advertiser’s own site.
OpenAI described the expansion as building advertising for the way people actually use the product. The company said ChatGPT now reaches 1.2 billion people a week, which it calls the world’s largest AI-native consumer platform, and it tied the new ad features to a set of advertising principles meant to reassure brands about where their messages appear.
The measurement piece is the part advertisers have been waiting for. Digiday reported the same day that ChatGPT budgets have been stuck at test level precisely because the numbers were missing. Marketers are willing to experiment with a new channel, but few will move meaningful money into one they cannot measure against the rest of their media. A pixel that reports conversions closes that gap, at least on paper.
OpenAI is pushing advertising even as it pushes other things. The company has spent the same period at its developer conference promoting new agent products and a higher-priced subscription tier, and it has signaled that advertising will sit alongside those businesses rather than replace them. The pitch is that subscription revenue pays for the model, while advertising pays for the reach.
OpenAI’s approach is to keep the ads small and tied to the conversation, so the product still feels like an assistant rather than a billboard. The company has argued that advertising helps it extend the product to people who cannot or will not pay, and that the two revenue lines reinforce each other. That argument will be tested by users who remember ChatGPT before there was any ad at all, and who will notice the moment the balance tips.
The timing matters because of what is happening around it. OpenAI is growing fast, and growth that relies on selling access to a model has limits. Advertising offers a second revenue line that scales with usage rather than with the number of people willing to pay a monthly fee. The company’s rivals are watching the same math, and the question hanging over ChatGPT is whether a conversational interface can become a place where advertising feels natural rather than intrusive.
That question is not settled. A chat window is a different surface from a search results page or a social feed. Users come to ChatGPT to get an answer, and an ad inserted into that exchange has to be relevant enough to justify the interruption. OpenAI’s answer is to keep the ads tied to the conversation and to give advertisers brand-suitability testing that lets them see, before they commit, whether their message will sit comfortably.
The broader contest is with the two companies that already know how to make advertising pay next to an AI product. Google and Meta run mature ad machines, and both are folding AI into them. OpenAI’s advantage is that it owns the assistant people talk to, a surface with a kind of attention that search engines no longer command so completely. The new formats and measurement are the company’s attempt to turn that attention into something advertisers can buy with confidence, the way they buy search and social today.
For advertisers, the appeal is straightforward. ChatGPT holds attention in a way few channels do, and the audience skews toward people actively seeking information rather than passively scrolling. That is the kind of attention marketers pay a premium for, provided they can measure what it returns. The new tools are aimed squarely at that proviso.
What remains to be seen is whether the measurement will hold up under real spending. A pixel is only as good as the attribution it produces, and advertisers will test the numbers against their own records before they trust them. If the figures survive, ChatGPT becomes a serious contender for budgets that currently go to search and social. If they do not, the ad business stays a curiosity, a promising surface that never quite pays for itself. OpenAI has now given buyers the instruments to find out.
