Search engines are built and optimized for people who don’t have the time or attention needed to scan an entire web page. But as more and more people use AI chatbots to search the web and perform tasks, the idea is that the internet’s current infrastructure should be updated to accommodate AI instead, as these bots can read and process much of the information.
Andrei Styskin, the former head of search, AI and cloud at Russian search giant Yandex, and German AI scientist Matthias Petri are working to solve that problem with their new startup, Keynable. The company recently came out of stealth by raising $26 million in seed funding. Accel led the funding round, with participation from Conviction Partners and some business angels.
From Styskin’s perspective, AI chatbots tend to perform much better if they can base their responses on the source document. “This actually creates a new flywheel that is different from what Google has learned from human behavior,” he told TechCrunch.
Keenable has built a web search index of over 100 billion documents and says its API is already used in production by several AI labs and inference providers for both training and runtime. The startup didn’t reveal who its customers are, but it recently partnered with voice AI company Gradium to support live information retrieval.
Drawing on his 20 years of experience building search at Yandex and Amazon, Styskin explained how Keenable’s product differs from enterprise search solutions that can prove broken and prohibitively expensive at web scale. “If we don’t fine-tune the index structure for specific tasks, the cost of servicing and scanning the entire Internet will be huge due to its volume. So we need to innovate ways to narrow down the search space based on a query very quickly, and this is what we are proposing,” he said.
According to Zhenya Loginov, an Accel partner who led the investment, AI players have few options when it comes to web-scale search infrastructure, especially as Google and Microsoft have taken steps to shut down their existing search APIs to avoid cannibalization. Instead, the tech giant is opting for a more holistic approach and is being selective about its partners.
For Styskin, these decisions confirmed the opportunities he saw while at Amazon, working with Petri on web search infrastructure for AI applications such as Alexa. After seeing data from Cloudflare showing that AI crawlers were responsible for an increasing share of search volume, he began to realize there was an opportunity to develop a web search infrastructure built with AI in mind.
Armed with that experience, Styskin leveraged his network to hire several former colleagues for his new startup. The company is also building its own search functionality. This includes the upcoming Web Query Language, which combines information from different web sources to help AI systems answer questions, even if they don’t contain the complete answer.
Cost becomes an important part of the equation. Steiskin said that while convincing people to move away from Google in search is “very difficult”, the innovator’s dilemma means the US giant could potentially “win” in agent queries, and smaller companies like Keynable could innovate and provide more cost-effective solutions to AI companies.
Still, there are real costs to building a large search index. “Don’t ask. It’s painfully expensive,” he said. But he said the startup is doing its best to keep costs down and keep pace. The company, which has a team of 15 engineering staff in the U.S. and Europe, plans to use the new cash to double its headcount by the end of the year and accelerate its market development efforts.
More steps are needed for the startup to achieve its dream of becoming “the next Google for AI agents.” Other players are entering this space, including Brave and Exa. And Google itself is overhauling its search experience for the AI era. But this wide-ranging motion shows that Keenable’s beliefs are shared at Googleplex. Whether it’s humans or agents, the days of the “10 blue links” may be coming to an end.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
