Taboola has opened the advertising system behind its DeeperDive AI answer engine to companies running conversational AI products, allowing chatbot and virtual assistant providers to sell advertising against user queries.
The platform lets external AI services use the same monetisation system that supports DeeperDive, Taboola's generative AI product for publishers. DeeperDive turns editorial content into a conversational format and places adverts within AI-generated answer pages.
Taboola is pitching the system to a growing group of AI developers that have attracted large audiences but are still searching for stable income beyond subscriptions. Under the model, an AI service can place an advert alongside an answer generated in response to a user question, with the advert selected to match the apparent intent behind that query.
In one example, a user asking about buying a home would be shown a mortgage advert. The aim is to let AI providers turn conversational interactions into advertising revenue without building their own commercial infrastructure.
Monetisation model
Advertising demand comes through Realise, Taboola's performance advertising platform. Realise connects to tens of thousands of advertisers through direct relationships, supplying the adverts that appear within DeeperDive and potentially within third-party AI services.
That extends a model Taboola already uses with publishers. DeeperDive has been embedded across a number of large news and media groups, where it gives readers an AI interface built on trusted editorial material and a way to ask follow-up questions on stories.
For publishers, the product is designed to create another advertising placement tied to user engagement with AI-generated answers. For AI developers beyond publishing, the same model offers a way to commercialise traffic that might otherwise be hard to monetise, particularly if users are unwilling to pay for multiple standalone AI subscriptions.
DeeperDive generates tens of millions of AI answers each month for more than 7 million users, according to Taboola. The system combines large language models, retrieval tools and what Taboola describes as its proprietary intent graph, which it uses to assess user behaviour and match commercial messages to likely interests.
The infrastructure also runs on NVIDIA accelerated computing, which helps support large volumes of AI responses with low latency. That matters for conversational products, where delays in serving answers or ads can hurt the user experience and reduce the chance of converting commercial intent into a click or sale.
Search for revenue
The announcement comes as AI companies continue to test different business models. Many consumer AI services have gained traction quickly, but the market has yet to settle on how those services will reliably make money at scale, especially where free access remains central to user growth.
Advertising has emerged as one possible answer, though it brings familiar concerns around relevance, user trust and the balance between utility and commercial interruption. By offering a ready-made system, Taboola is seeking to position itself between AI providers and advertisers in much the same way it has long done across publisher websites.
The strategy also reflects a broader shift in how internet platforms are thinking about user discovery and engagement. If users increasingly move from browsing web pages and search results to asking questions in chat interfaces, advertising businesses will need ways to fit commercial content into those exchanges without making the experience feel forced.
Taboola says its existing position across publisher sites and device manufacturers gives it the data and scale to support that transition. Its network reaches about 600 million daily active users through publisher and original equipment manufacturer relationships including NBC News, Yahoo, Samsung and Xiaomi.
Adam Singolda, Chief Executive Officer at Taboola, framed the product as an attempt to address the economics of conversational AI. "Today's news is about helping build the economic layer of the AI internet," he said.
He said many AI developers still lacked durable ways to support their products. "The future of the internet will be increasingly conversational and agentic. While AI companies have made incredible progress building products consumers love, many are still looking for sustainable business models. Consumers won't subscribe to every AI service they use, creating a significant opportunity for advertising and commerce to help fund innovation. DeeperDive has proven that AI experiences can successfully connect users with trusted content and relevant commercial ads. We're now bringing that monetisation infrastructure to the broader AI ecosystem, helping AI applications, agents, and LLM-powered services grow while creating value for users, advertisers, and publishers," Singolda said.