Accel-Backed Keenable is Indexing the Web for AI Agents.
Image Credits:Keenable
The Changing Landscape of Search Engines: Cater to AI
Search engines have traditionally been designed for human users who often lack the time to read entire web pages. However, as AI chatbots gain traction for web searches and task execution, a shift in focus is necessary—one that prioritizes AI rather than human users. This is largely due to AI’s capability to process extensive data volumes more efficiently.
Introducing Keenable: A New Player in AI-Driven Search
Andrey Styskin, who previously spearheaded the search, AI, and cloud divisions at Yandex, alongside German AI scientist Matthias Petri, aims to tackle this challenge with their startup, Keenable. Recently emerging from stealth mode, Keenable has secured $26 million in seed funding, with Accel leading the round and contributions from Conviction Partners and various angels.
The Need for Grounded AI Responses
Styskin believes that AI chatbots provide more accurate responses when they can reference source documents. He explains, “This creates a different flywheel compared to what Google learned from human behavior.” Keenly aware of the evolving landscape, Keenable is developing a web search index comprising over 100 billion documents. Their API is already in active use across several AI labs and inference providers, both during training and runtime.
A Different Approach to Search Technology
Styskin draws on two decades of experience in search at Yandex and Amazon to differentiate Keenable’s offerings from existing enterprise search solutions that are often expensive at a web scale. He notes that without tailored index structures, costs can skyrocket due to the vast amount of data involved. “It’s imperative to innovate in how quickly you can refine the search space based on your query,” he adds.
Competing with Industry Giants
Accel partner Zhenya Loginov, who led the recent investment, points out that AI companies face limited choices regarding web-scale search infrastructure. Google and Microsoft have been shutting down existing search APIs to prevent internal competition, leading them towards a more consolidated partnership model. Such industry dynamics reaffirm the opportunity Styskin recognized while collaborating with Petri at Amazon, especially in developing web search infrastructure for AI applications like Alexa.
Building a Robust AI Infrastructure
Styskin leveraged his network to recruit former colleagues who share his vision for Keenable, which is also focused on developing proprietary data retrieval capabilities. One notable product in the pipeline is WebQueryLanguage, aimed at empowering AI systems to respond to inquiries by aggregating information from various web sources, even when individual sources lack complete answers.
Cost Considerations in Indexing
A key factor for Keenable will be managing costs. Styskin acknowledges the challenge of persuading users to shift away from Google for searches. However, he believes that due to the “innovators’ dilemma,” the tech giant is not invincible, particularly for agentic queries. He argues that a smaller enterprise like Keenable can offer more innovative and cost-effective solutions.
Nevertheless, the financial burden of constructing a large-scale search index is significant. “Don’t ask—it’s painfully expensive,” Styskin admits. Nonetheless, his startup is working diligently to keep costs manageable while pacing its growth. With a current engineering staff of 15 across the U.S. and Europe, Keenable aims to double its team by the year’s end to facilitate its go-to-market strategy.
The Road Ahead for Keenable
While Keenable aspires to become “the next Google for AI agents,” the path is fraught with challenges. Competitors like Brave and Exa are vying for a position in this space, and tech giants like Google are overhauling their search experiences for an AI-first world. This broader evolution highlights a shared belief in the industry: the traditional era of “ten blue links” may be nearing its conclusion, be it for humans or AI agents.
In summary, the rapid shift towards AI-driven search emphasizes the need for a revamped infrastructure capable of catering to both AI and human users. Keenable’s innovative approach, extensive indexing, and unique retrieval capabilities position it well for this new era of search, though significant hurdles remain. The startup exemplifies the potential for disruption in an industry long dominated by major players, hinting at a future where AI-driven searches become increasingly predominant.
Thanks for reading. Please let us know your thoughts and ideas in the comment section down below.
Source link
#Accelbacked #Keenable #indexing #web #agents
About The Author
