Until now, ads in ChatGPT looked like a controlled experiment: one brand, one slot, an audience nobody knew how to measure. That just changed.
OpenAI is testing a multi-advertiser format with real-time auctions, daily budget models, and campaign management tools that look increasingly like Google Ads. MarTech.org recently detailed the technical specs. But the real question is not what OpenAI is testing. It is what they are building. The short answer: an advertising platform that is taking shape much faster than most expected.
Auctions, budgets, and the Google Ads playbook
OpenAI has moved to a second-price auction model, identical in principle to Google Ads. The highest bidder wins the placement but pays a marginal price above the next bid. If you have been running PPC for a few years, the mechanism is familiar. And that is exactly what makes it significant: OpenAI is not inventing a new pricing model. They are adopting the one that already works at scale.
Alongside auctions, they have added features that make no sense in a beta test but make perfect sense in a platform preparing for broad launch: automatic lifetime-to-daily budget conversion, CPM-to-CPC campaign cloning, and bulk editing tools. Anyone who has worked in Google Ads Editor or Meta Ads Manager recognizes the pattern.
The practical result? Real competition for placements. When multiple advertisers bid on the same conversational intent, prices rise. And budget predictability drops if you do not have a bidding strategy adapted to this new context. It is a dynamic we already know from AI-driven auctions in Google Ads, but applied on completely different terrain: conversations, not searches.
From 4 markets to 9. This is no longer a test.
Initially, ChatGPT ads were limited to the US, Canada, Australia, and New Zealand. Now the UK, Japan, South Korea, Brazil, and Mexico have been added. From 4 markets to 9 in a single iteration, a jump of over 100% in geographic coverage.
When a platform makes this kind of expansion, the signal is clear: they have enough data to know the model works. They have enough advertiser demand to justify the infrastructure. And they have enough confidence in placement quality to expose it to markets with very different advertising expectations. Japan and Brazil have advertising cultures that are nothing like the US, which suggests the format holds up across varied contexts.
If you work with clients across European or global markets, the relevance is immediate. Even if you only target a local market, the direction matters: AI chat is becoming an advertising channel, not just an information channel. We discussed the concept of conversational advertising a few months ago. What we are seeing now is that idea being built into actual ad infrastructure with real auctions.
The broader context matters too. Google is testing increasingly aggressive ad formats in AI Mode and AI Overviews. Meta is investing heavily in advertising on Threads and in AI experiences within WhatsApp. Now OpenAI enters direct competition for advertising budgets with a platform built on conversation. For media buying teams, this means one more channel to monitor, one more set of metrics to learn, and one more competitor for the same audience attention. Fragmentation is not new, but the speed at which AI-based channels are emerging is unprecedented.
Why the multi-advertiser format changes the equation
In a single-advertiser-per-slot model, OpenAI controls the price directly. In a multi-advertiser auction model, the market sets the price. The difference is massive.
With a single advertiser, OpenAI could test whether users accept ads. With multiple advertisers, they are testing something entirely different: how much brands are willing to pay for the attention of someone actively asking about a real problem. And the history of Google Ads tells us that on purchase-intent queries, the answer is consistently high.
Think about the conversational context. Someone asks ChatGPT: "What is a good business laptop under $1,500?" The response comes with a recommendation. And now, alongside that response, a sponsored placement from a relevant manufacturer appears. The intent is clear, the context is precise, and the user is already in evaluation mode. From an advertising perspective, this is a high-conversion moment.
What this means for your PPC team
Budgets need rethinking. A new channel with real auctions means a cost per click you do not know yet. Testing with small budgets (around 10-15% of your experimentation budget) is rational in markets where the ads are available. This is not the moment to move significant budget. It is the moment to learn how pricing works, what types of queries generate placements, and what click-through rates look like in a conversational context.
Creatives change fundamentally. ChatGPT ads appear in conversational context, between the answers a user receives to real questions. This is not a banner on a website, not an ad in the SERP. The message needs to be useful within the conversation, or it will be ignored. Aggressive sales copy performs poorly here. Helpful, relevant recommendations perform much better.
Measurement becomes a puzzle. How do you attribute a conversion from ChatGPT? Cookies do not work the same way as in a browser. The user journey is completely different from search: someone asks, gets a response, sees an ad, maybe clicks, maybe continues the conversation. Attribution models in GA4 are not natively ready for this. Your analytics team will need new configurations and, likely, several months of data before drawing solid conclusions.
Google Ads competition redistributes. Advertisers testing ChatGPT ads may reduce pressure on certain keywords in Google. Or they may increase total cost if both platforms bid on the same audience with similar intent. It depends on vertical, budget, and how aggressively your competitors move. Not every advertiser will benefit equally from a new channel, especially those who do not adapt their message to the context.
Preparation, not panic
ChatGPT is no longer testing ads. It is building an advertising platform with auctions, budgeting, and campaign management. It takes the logic you know from Google Ads and applies it to AI conversations. The transition from experiment to infrastructure happened faster than the industry expected.
You do not need to move budget today. You need to understand the mechanism, prepare your team for a new channel, and monitor when your market becomes eligible. Those who test and learn first will have a real advantage when competition intensifies. And it will.
