OpenAI’s Ad Push Gets a Smartly Boost

OpenAI has already said its ad business is moving from theory to rollout, with more than 600 advertisers and $100 million in annualized recurring ad revenue on the board. Now it has picked Smartly as its first creative adtech partner, a sign that ChatGPT ads are headed toward something more interactive than a static banner or a sponsored link.

The pitch is simple: if people are already asking ChatGPT for advice, brands want a way to answer inside that same flow. Smartly, which has spent years helping advertisers like Spotify and Uber tune campaigns in real time, wants to bring conversational ad formats into ChatGPT without breaking user trust.

OpenAI’s ad rollout is getting more specific

OpenAI started testing basic ads in February for users on its free tier and its $8-a-month Go tier. Those early placements were contextual, which means the system matched ads to the topic of the conversation. Someone comparing phones might see a Best Buy ad. Someone planning a trip might see Expedia.

That setup is useful, but it is still pretty close to old-school search advertising. The next step is more ambitious: ads that feel native to ChatGPT’s back-and-forth style. Smartly’s role is to help OpenAI and its advertisers make those placements more personalized and more responsive to what a user asks next.

OpenAI already has another adtech partner in Criteo, which helps brands place ads in ChatGPT. Smartly is different because it is focused on creative optimization, the part of the stack that changes what the ad says, how it looks, and how it reacts to performance data.

• OpenAI says ads will stay separate from organic answers
• Ads will be clearly labeled inside ChatGPT
• Users under 18 will not see ads
• Ads will not appear near politics or health topics
• OpenAI says advertisers will not get access to private user conversations

Why Smartly matters more than a normal ad partner

Smartly is a 13-year-old adtech company led by CEO Laura Desmond, a veteran of the media and advertising world. Its pitch to OpenAI is not just about placing ads, but about making them feel useful in a chat interface. That matters because the more an ad feels like a distraction, the faster users tune it out.

Desmond told Business Insider that conversational ads can do better by asking follow-up questions and refining recommendations in real time. That is a very different model from the usual one-shot ad impression. It also fits the way people already use ChatGPT: they ask, refine, compare, and ask again.

“The opportunity with conversational advertising is you can do more follow-ups, and you can ask again,” Laura Desmond said. “The experience for people will get way more relevant, way more personal, and hopefully be seen as a much better value exchange.”

Smartly has already tested this idea outside OpenAI. Desmond pointed to a Boots campaign in the UK, where a chatbot-style ad on Meta platforms like Instagram opened a new window and asked questions before offering gift recommendations. Smartly said that format was nearly five times as effective at driving sales as Meta’s basic ads.

That number matters because it shows what OpenAI is chasing. If ChatGPT can turn ad inventory into an interactive recommendation engine, it could sell a much more valuable placement than a generic sponsored result.

The money at stake is huge, but so are the constraints

OpenAI has a real shot at building a meaningful ad business. In January, Evercore ISI analyst Mark Mahaney said OpenAI could generate several billion dollars in ad revenue this year and as much as $25 billion by 2030. That is a huge range, but even the low end would make ads a serious part of OpenAI’s business mix.

At the same time, the company is working inside tight limits. ChatGPT does not have endless screen space for promotions, and OpenAI has said it will not embed ads directly into its answers. That means the company has to create ad formats that feel helpful without making the product feel crowded.

OpenAI also has to avoid the trust problem that haunts most ad-supported products. If users think their chats are being mined for targeting, the product gets less useful fast. OpenAI says it will keep conversations private from advertisers and will not sell user data to them. That is a strong promise, but it also narrows the kinds of targeting it can offer.

• OpenAI says it is on track for $100 million in annualized recurring ad revenue
• The company says it works with more than 600 advertisers
• Sensor Tower found more than 100 brands had advertised on ChatGPT in the first few weeks
• 44% of those brands were retail companies
• Smartly said its Boots conversational ad format was nearly 5x as effective as Meta’s basic ads

How ChatGPT ads compare with Google, Meta, and Anthropic

OpenAI is entering a market dominated by Google and Meta, two companies that have spent years perfecting ad targeting, measurement, and auction mechanics. That is a hard benchmark to meet because both companies own huge amounts of user attention and have mature ad systems built around it.

There is also a philosophical split among AI companies. Anthropic has rejected ads in its Claude chatbot, arguing that ads would undermine its mission to be helpful. Google takes a middle path: it shows ads in AI Overviews, but not in its Gemini chatbot. OpenAI is trying to thread the needle between those positions.

The real question is whether users will accept ads in a product they treat like a helper rather than a feed. That is why Smartly’s work matters. The company is not just helping OpenAI sell inventory. It is helping define what an ad can look like when the interface itself is a conversation.

If OpenAI gets this right, the format could become a template for other AI products that want ad revenue without turning into cluttered search pages. If it gets it wrong, ChatGPT could end up with the same problem that has hit every ad-supported platform before it: too much monetization, not enough trust.

What happens next

OpenAI’s ad business is still early, but the Smartly partnership tells us where it is headed. The company is moving from simple contextual placements toward interactive ads that ask questions, adapt in the moment, and try to feel like part of the conversation.

My read: the next 12 months will be less about scale and more about restraint. OpenAI will likely test a small number of high-intent categories such as retail, travel, and entertainment, then expand only if users do not react badly. If conversational ads feel useful, ChatGPT could become one of the most valuable new ad surfaces on the internet. If they feel intrusive, OpenAI may have to slow the rollout and protect the product before the revenue.

The key thing to watch is simple: does OpenAI keep ads feeling like answers, or do they start feeling like interruptions?