Keenable Rebuilds Web Indexing to Serve AI Agents

Search engines have long been designed to suit the needs of the human user who lacks the time or patience to read entire web pages, settling instead for a concise list of links. But as people increasingly rely on smart chatbots to search and accomplish tasks, a growing view has taken hold that the internet's current infrastructure needs to be redesigned to suit machines rather than humans—especially since these systems are capable of reading and processing vast amounts of information simultaneously.

Who Is the Founding Team?

A startup called Keenable aims to address this gap. It was founded by Andrey Styskin, who previously led the search, artificial intelligence, and cloud computing division at Russian search giant Yandex, alongside German AI scientist Matthias Petri. The company recently emerged from stealth mode after raising $26 million in seed funding led by Accel, with participation from Conviction Partners and a group of angel investors.

Styskin believes that smart chatbots deliver far better results when their answers are grounded in reliable source documents, arguing that this creates a new self-improving cycle fundamentally different from the one Google built based on user behavior.

A Massive Web Index

The company says it has built a web search index comprising more than 100 billion documents, and that its API is already in use in production environments at a number of AI labs and inference service providers, whether during training or actual operation. While it declined to reveal the identity of its clients, it recently announced a partnership with Gradium, a company specializing in voice AI, to support real-time information retrieval.

What Sets Keenable Apart Technically?

Styskin draws on twenty years of experience building search systems at Yandex and Amazon, explaining how his company's products differ from enterprise search solutions that may collapse or become prohibitively expensive at web scale. He notes that scanning the entire internet consumes enormous resources unless indexing structures are precisely tuned for each individual task—and this is where the innovation lies: rapidly narrowing the search space based on the query itself.

The company is developing its own retrieval capabilities, including an upcoming product called Web Query Language, which allows AI systems to answer questions by combining information from multiple web sources, even when none of them contains the complete answer.

An Opportunity Created by the Giants' Restrictions

According to Jenia Loginov, the Accel partner who led the investment, players in the AI field have only limited options when it comes to web-scale search infrastructure, particularly as Google and Microsoft move to restrict their own search interfaces to avoid competing with their core products, favoring a more integrated and selective approach in choosing partners.

Styskin sees these decisions as confirmation of the opportunity he spotted during his work at Amazon on web search infrastructure for applications like Alexa, when he noticed from Cloudflare data that AI crawlers were capturing a growing share of search volume.

The Challenge of Cost and Competition

Styskin acknowledges that building a giant search index is extremely costly, but points out that the company is keen to control expenses and advance at a measured pace. Its team currently includes about 15 engineers distributed between the United States and Europe, and it plans to use the new funding to double its headcount by year's end to build its go-to-market strategy.

Despite the difficulty of convincing users to abandon Google, the founder believes the American giant can be challenged on queries tailored to AI agents, where a smaller company can innovate and offer a lower-cost solution. The most prominent challenges facing Keenable:

• The exorbitant cost of building and operating a massive web index.
• Competition from emerging players like Brave and Exa.
• Google reinventing itself for the search experience in the AI era.

The road ahead remains long for the company to realize its dream of becoming "the next Google for AI agents," yet this accelerating momentum points to a conviction shared even by Google itself: whether search is aimed at humans or at agents, the era of the "ten blue links" may be drawing to a close.

✦ بقلم فريق دروب أيديا