On October 21, OpenAI said it was expanding ChatGPT Ads to 26 more countries, bringing the total to 89. The same announcement said more than 100,000 advertisers have run campaigns since ads launched a little over eight months ago. The new markets include 14 more in Latin America, including Argentina, Chile, Colombia, and Peru, plus Nigeria, South Africa, Serbia, Ukraine, Cambodia, Bangladesh, and Pakistan.
Two details matter more than the country count. First, you can buy these yourself. OpenAI says advertisers can go through a technology partner or use self-serve through Ads Manager. Second, ads appear only to people on the Free and Go plans. Plus, Pro and Enterprise stay ad-free.
Google is moving in the same direction. Search Engine Land reported in mid-October that Google began testing ads in AI Mode for users who connect their apps through Personal Intelligence, with that connected-app information available to personalize the ads.
So the answer box is becoming an ad unit. On both sides.
Should you buy them?
I am going to disappoint you. I do not know, and neither does anyone writing a thousand words about it this week.
Here is what we do know. The inventory is new, which usually means it is cheap relative to what it will cost in a year. OpenAI didn’t publish minimum spends or pricing in that announcement. The audience is specifically people who did not pay for a subscription, which says something about their willingness to spend, though not everything.
What nobody can tell you yet is whether a person halfway through a conversation about which contractor to hire behaves like a search clicker, a social scroller, or something new. The published advertiser count tells you how many businesses tried it. It does not tell you how many are still there.
What decides where to advertise your business?
Not the channel. What you already know.
If you cannot say what a customer is worth to you over a year, and what you currently pay to get one, then a new channel is not an opportunity. It is a faster way to spend money without learning anything. You will not be able to tell whether ChatGPT Ads worked, because you have no baseline to compare it to.
If you can say those two numbers, the decision gets simple and boring. You have a cost per acquisition you can live with. You run a small, bounded test. You find out in six weeks whether this channel clears that bar. If it does, you scale it. If it doesn’t, you stop, and the test costs you a known amount.
That is the whole discipline. It is not sophisticated. It is just unglamorous enough that most owners skip it and then argue about platforms instead.
One caveat about being early
Being early on a channel offers a real advantage, and I have taken it more than once in forty years. Cheap inventory before the auction fills up is a genuine edge.
But early also means no benchmarks, thin targeting, changing formats, and measurement that is still being built. OpenAI said future phases will bring additional formats, objectives, buying options, and measurement capabilities. Read that as: measurement is not finished. Being early is a bet you place with money you can afford to lose while you learn, not with the budget that has to produce leads this quarter.
If you want someone to build and run that test rather than guess at it, that is the work a PPC agency does. And if the underlying question is less which platform and more how to tell strategy from tactics in the first place, that is the argument I make at length in The Art of Marketing.