Analysis
Keenable, a web-search infrastructure startup built specifically for AI agents rather than human browsers, raised a $26 million seed round led by Accel, TechCrunch reported, with Conviction Partners and a group of angel investors also participating. The company exited stealth with this announcement in August 2026.
The Founding Team's Search Pedigree
Co-founder Andrey Styskin previously led search, AI and cloud infrastructure at Yandex, Russia's largest search engine, before working at Amazon; co-founder Matthias Petri worked alongside him there on the search infrastructure powering Alexa. That combined background -- building search at both a major search-engine company and a voice-AI product used by hundreds of millions of people -- is the core credibility behind Keenable's pitch that agent-facing search requires fundamentally different index architecture than the general web search Google and Bing built for decades.
“Keenable's roadmap includes a forthcoming Web Query Language product meant to let agents query the index more precisely than natural-language search terms allow.”
- Keenable -- $26M seed, Accel-led, 100B+ document index built specifically for AI-agent retrieval
- Brave Search, Exa -- competing search-API providers also targeting AI developers and agent builders
- Google -- the default general-purpose search infrastructure Keenable is positioning against for agent-specific use cases
Why Agent Search Is a Different Problem
Styskin's framing of the technical bet is direct: "If you do not fine-tune your index structures for a specific task, the cost of serving and scanning the whole internet is enormous." The argument is that human search queries and AI agent queries look meaningfully different -- agents often need live, verifiable information retrieval chained into multi-step reasoning, not a ranked list of blue links optimized for click-through -- and that a general-purpose index built for human search behavior carries cost and latency overhead that doesn't map well onto how agents actually consume retrieved information. Keenable's roadmap includes a forthcoming Web Query Language product meant to let agents query the index more precisely than natural-language search terms allow.
Pulse has tracked Amazon's broader AI and infrastructure moves separately, relevant given Keenable's founders built their search chops on Amazon's Alexa team.
The Competitive Reality
The honest challenge Keenable faces is that Brave Search and Exa have both been building toward the same agent-search thesis for longer, and Google -- despite being positioned as the incumbent Keenable is differentiating against -- has vastly more crawling infrastructure, index freshness, and compute to draw on if it decides to build a dedicated agent-search product of its own. A $26 million seed is enough to prove the technical thesis with early API customers, but building and maintaining a 100 billion-plus document index competitive with Google's crawl infrastructure over the long run will require substantially more capital in future rounds.
What to Watch
The next real signal for Keenable is which AI labs or inference providers, if any, publicly adopt its API as a production dependency -- a single major AI lab standardizing on Keenable's index would validate the agent-search thesis in a way seed-stage funding announcements alone can't.