Guru Pitches Verified Workplace Knowledge Delivery to AI Tools

What changed

A sponsor message published on October 27 describes Guru, a knowledge-management product that turns resolved workplace discussions and tickets into drafts, verifies content using plain-language rules, and delivers approved information to AI tools through MCP. It claims 80% lower token use, 16% fewer repeat questions, and 96% of searches ending with an AI answer, though those figures are attributed only to unspecified teams.

Why it matters

The useful idea here is less dramatic than the headline suggests: an AI system is only as dependable as the knowledge it can retrieve. Guru’s pitch is to keep outdated answers out of that retrieval layer and turn repeated workplace explanations into reusable documentation.

That could reduce the familiar loop in which someone answers a question in Slack, the answer disappears into the scroll, and the next person asks again. But the headline’s references to Anthropic, OpenAI and the “end of the internet” are not supported by the message itself. The reported metrics should therefore be treated as product claims, not established results.