Search engines have long been designed with humans in mind—people who don’t have the time or attention span to sift through entire webpages. But as AI chatbots become the go-to tool for web searches and everyday tasks, a growing number of experts argue that the internet’s current architecture needs a rethink. After all, these bots can process far larger volumes of information than any human ever could. Andrey Styskin, who once led Russian search giant Yandex’s search, AI, and cloud divisions, along with German AI researcher Matthias Petri, are tackling this challenge head-on with their new venture, Keenable. The startup recently emerged from stealth with $26 million in seed funding, led by Accel, with participation from Conviction Partners and several angel investors.

From Styskin’s viewpoint, AI chatbots perform significantly better when they can anchor their responses in source documents. Keenable claims to have built a web search index spanning over 100 billion documents, and its API is already being used in production by several AI labs and inference providers, both during training and at runtime. While the startup hasn’t disclosed its clients, it has recently partnered with voice AI company Gradium to enable real-time information retrieval.

Drawing on two decades of experience building search systems at Yandex and Amazon, Styskin explains what sets Keenable apart from enterprise search solutions, which often struggle and become prohibitively expensive at web scale. “If you do not fine-tune your index structures for a specific task, the cost of serving and scanning the whole internet is enormous because of the volume. That’s why you need to innovate on how you can narrow the search space based on your query very fast. This is what we are bringing to the table,” he said.

According to Accel partner Zhenya Loginov, who led the investment, AI companies have limited options when it comes to web-scale search infrastructure. This is especially true as Google and Microsoft have moved to shut down their existing search APIs to avoid cannibalizing their core businesses. Instead, the tech giants are favoring a more bundled approach and being selective about their partners. For Styskin, these developments validated the opportunity he first spotted while at Amazon, where he worked with Petri on web search infrastructure for AI applications like Alexa. He had noticed Cloudflare data showing that AI crawlers were accounting for a growing share of search volume, and he began to see the potential for building web search infrastructure specifically designed with AI in mind.

Armed with that insight, Styskin tapped his professional network to recruit a handful of former colleagues for the new startup, which is also developing proprietary retrieval capabilities. One upcoming product, WebQueryLanguage, is designed to help AI systems answer questions by synthesizing information from multiple web sources, even when no single source contains the complete answer.

Cost will be a critical factor. Styskin acknowledges that while it’s “extremely hard” to persuade people to abandon Google for search, the innovators’ dilemma means the U.S. giant is potentially “beatable” on agentic queries. A smaller company like Keenable, he argues, can innovate and deliver a more cost-effective solution to AI firms. Still, the expense of building a massive search index is daunting. “Don’t ask—it is painfully expensive,” he said. But the startup is working to keep costs under control and pace itself accordingly.

With a team of 15 engineers spread across the U.S. and Europe, Keenable plans to use the fresh funding to double its headcount by the end of the year and build out its go-to-market strategy. There’s still a long road ahead before the company can realize its ambition of becoming “the next Google for AI agents.” Other players, like Brave and Exa, have entered the space, and Google itself is revamping its search experience for the AI era. But this broader momentum suggests Keenable’s conviction is shared even at the Googleplex: whether the user is human or an agent, the era of the “ten blue links” may be drawing to a close.