Let me say something that might sound counterintuitive for a founder in this space: OpenAI figuring out ads is the best thing that could have happened to us.
Not because we are competing with them. We are not. But because when the largest AI company in the world (famously resistant to advertising) quietly builds a multimillion dollar ads business, it answers the one question every investor and every AI publisher has been sitting on:
Is there actually money in monetising AI conversations? Yes. Definitively yes. And now everyone knows it.
Here is the problem. There are now hundreds of AI apps: astrology platforms, health assistants, finance tools, productivity apps, coding helpers. All sitting on enormous volumes of high-intent user queries. They watched OpenAI do it. They want in. And there is absolutely no infrastructure for them to plug into.
That is the gap BidTurns exists to close.
AI conversations carry the richest intent signal ever seen in advertising
I spent close to a decade in advertising at The Trade Desk, Adobe, and InMobi. I have sat on both sides of the market. I have watched publishers struggle to monetise and advertisers struggle to reach the right people at the right moment.
The thing that struck me when conversational AI started scaling was not the technology. It was the signal.
When someone opens an AI app and types "I'm three months postpartum and I can't sleep. What actually helps?" that is not a search query. That is a person telling you exactly where they are in their life, what they need, and approximately what they are willing to spend to solve it.
The intent signal in a conversational AI query is richer than anything the advertising industry has ever had access to. Richer than search. Richer than social. And it is being generated at a scale that is growing by the day.
I could not find anyone building the infrastructure to act on it cleanly. So my co-founder Akshay and I started BidTurns.
OpenAI built a walled garden. The other 500 AI apps have nothing.
Here is where I want to be precise, because this is the part that gets misunderstood.
Yes, OpenAI is building ads. Yes, Perplexity has a sponsored answers product. Yes, a handful of the largest AI platforms are figuring out their own monetisation. But here is what they are doing: they are building walled gardens.
OpenAI monetises OpenAI traffic. Perplexity monetises Perplexity traffic. Each platform is building its own demand relationships, its own auction, its own format. Independently, from scratch, for their own inventory only.
This is exactly how the web looked before ad networks existed. Every publisher was cutting individual deals with individual advertisers. Inefficient, unscalable, and it left most of the market completely locked out.
Ad networks fixed that for the web. Programmatic fixed it for display. Real-time bidding fixed it for mobile. Nobody has fixed it for conversational AI.
The hundreds of AI apps that are not OpenAI (the ones with 500,000 users, 2 million users, 10 million users) have no way to access advertiser demand at scale. They cannot afford to hire a sales team. They do not have engineering resources to build an auction from scratch. They are stuck.
That is who we are building for.
The hard part isn't the auction. It's knowing when to fire.
I want to be honest about something. Building this is genuinely difficult. If it were easy, it would already exist.
The challenge is not the auction mechanics. Programmatic infrastructure is well understood. The challenge is what has to happen before the auction: the intent layer.
In display advertising, you target a person based on who they are: demographic, browsing history, inferred interests. The context is secondary. In conversational AI, the person is almost irrelevant. We do not know who they are. That is deliberate. Privacy is non-negotiable in this medium.
What we have instead is an extraordinarily precise picture of what they want, right now, in this conversation. Acting on that correctly requires understanding intent at the turn level: not just what the user asked last, but how the signal has been building across the whole session.
Has enough intent accumulated across turns to warrant a recommendation or is the signal still too weak? As the conversation deepens, which advertiser fits where this is heading? And when the moment finally arrives, what tone makes the recommendation feel earned rather than jarring?
Get this wrong and you destroy the user experience. Get it wrong repeatedly and you poison the well for every AI app trying to monetise. This is why the mediation layer cannot be bolted together from existing parts. It has to be built from scratch, specifically for this medium.
The founding window is open. It won't be for long.
Every infrastructure category has a founding window a period where the standards are being set, the first integrations are going live, and the network effects beginning to accumulate.
For programmatic display, that window was roughly 2008 to 2012. The companies that built during that period did not just build businesses. They became the pipes that the entire industry ran on for the next decade.
Conversational AI monetisation is in that window right now. Every AI app founder I speak to has the same question: how do we monetise without compromising the experience?
We have the answer. One SDK. Real-time intent detection. Native format delivery. A live auction connecting publisher inventory to advertiser demand. Zero PII. Full publisher controls. And the advertisers want in. Reaching someone at the precise moment they are deciding, without cookies or guesswork, is what performance marketers have been hunting for since the death of the third-party cookie.
Both sides of the market are ready. The infrastructure to connect them is what has been missing.
The standards set now will define AI advertising for a decade
I am not going to pretend this is purely altruistic. We are building a business.
But I genuinely believe the way this gets built matters. The standards set now around transparency, user experience, and what constitutes an acceptable ad in a conversational context will shape how AI advertising works for the next decade.
Done badly, AI advertising will erode the trust that makes these platforms valuable in the first place. Done well, it creates a sustainable revenue model that lets AI apps stay free, keeps advertisers honest, and gives users recommendations that are actually useful.
That is the version we are trying to build.