OpenAI rolled out a slate of updates to ChatGPT Ads Manager in late July, Search Engine Land reports. Advertisers can now create conversion-optimized campaigns by selecting a Conversions objective; OpenAI says campaigns "automatically optimize toward clicks that are more likely to generate conversions while continuing to charge on a CPC basis." The release also adds geographic exclusions, mobile measurement integrations with AppsFlyer and Adjust, automatic advanced matching for website conversions using hashed customer data, and asynchronous bulk operations in the Ads API for creating and editing campaigns at scale. Product feed ads got refreshed cards that display prices and star ratings.

Budgets changed too. Starting July 27, a daily budget in ChatGPT Ads stopped being a fixed daily cap and became an average over a rolling seven-day period, so a campaign can spend more on one day and less on another as long as the week balances out, per OpenAI's email to advertisers reported by Search Engine Roundtable. It is the same pacing model Google and Meta use.

The cadence is the context. OpenAI opened self-serve ads to all US advertisers earlier in July after dropping its spend minimum, added custom audiences and new ad formats the week before that, and is now filling in bidding, budget pacing, measurement, and bulk tooling within weeks of each other.

The takeaway for brands: ads inside ChatGPT are getting real machinery, and plenty of marketers will use it. Worth keeping straight is what that machinery does and does not touch. The answer beside the ad is still written by the model from its own reading of the market, and buyers see the two together. An ad can place you next to that judgment; it cannot change what the model actually believes about you. Brands that treat the ad slot as the whole ChatGPT story will miss the half of the screen where the model speaks its own mind.