Read the answer before you buy the ad.
ChatGPT Ads put a sponsored card directly under the answer to your buyer’s question, and OpenAI says ads don’t influence what that answer says. So the answer decides how hard your ad has to work. We use our GEO data to see what ChatGPT, Gemini or other chatbots already tell buyers about you and your competitors, then put budget where the shortlist is still open: where you’re one name among several, or missing entirely.
The answer above your ad goes one of three ways, and each gets a different plan:
Where ChatGPT already picks you: prove the ad adds something.
An ad under an answer that already recommends you is the brand search question all over again. We test those topics against a holdout and fund them only if the ads bring in customers you wouldn’t have won anyway.
Where you’re one of six, or missing: give the buyer a reason.
Ask ChatGPT for the best POS system for a small coffee shop in the Southwest and you’ll get a shortlist. If you’re one name on it, or not on it at all, the ad makes the case the answer didn’t, such as a rating from coffee shop owners or the number of cafés already on the system. We lead with proof built for the buyer’s exact situation.
See how we track recommendation rate
Where ChatGPT steers buyers away: fix the reason or skip it.
If the answer tells a buyer you’re the wrong choice, an ad underneath won’t win them back. When the model is right, we exclude the topic. When its reason is out of date, we fix the sources it read.
More on GEO
We choose the conversations.
ChatGPT Ads have no keywords. We build ad groups around topics where you’re one of many or missing from the answer, and exclude the ones where the model steers buyers elsewhere.
We wire the signal first.
The OpenAI pixel and Conversions API go live before launch, with the conversion event chosen up front, because a ChatGPT campaign’s objective is locked once it runs.
We lead with proof.
One headline, one line of copy, carrying the proof point the answer left out: ratings, customer counts and results in the buyer’s own vertical.
We measure past the platform.
UTMs into your analytics and CRM, view-through kept separate from clicks, and results benchmarked against non-brand search. The topics that convert go back to our GEO team as prompts worth winning organically.
We put ChatGPT budget where the answer still leaves your buyer choosing.
Early results are mixed.
That’s when a small test teaches the most. OpenAI says ChatGPT Ads passed a $1 billion annualized run rate within 200 days of launch, with tens of thousands of advertisers already buying. Some agencies say it isn’t performing for them yet. Both are what a young ad platform looks like: real demand, beta tools and reporting that’s still filling in. We cap the test budget, set the bar for success before launch, and tell you whether to scale or wait.
