Glean: Product Launch

The leadership read

Glean's move from retrieval to MCP server is not a feature addition; it is a redefinition of where the product sits in the enterprise stack. Previously, Glean competed on search quality inside its own interface. Now it is positioning as the context layer for third-party agentic tools, meaning its stickiness no longer depends on users opening Glean directly. The spillover dynamic Emrecan Dogan described, where external harness usage pulls users back into Glean surfaces, inverts the normal integration logic: instead of Glean embedding into other products, other products are becoming a distribution mechanism for Glean's data layer. That is a meaningfully different retention and expansion model than the company was running before. This is one of twelve product-launch signals in agentic AI infrastructure we have tracked in the last 90 days. The most structurally comparable moves are Slack's launch of Slack Code (embedding Claude and Devin into team workflows) and Raft's persistent-agent collaboration platform, both of which bet on agents as the primary interface rather than a feature layer sitting beneath a human-facing UI. The consistent pattern: enterprise software vendors are racing to position their data or workflow layer as the canonical context source for a multi-agent environment, rather than the destination. Across companies reaching this stage of agentic platform build-out, the pattern consistently surfaces demand for product leadership at the seam between API/developer ecosystems and enterprise GTM, commercial operators who can price and package context-as-a-service rather than seat-based SaaS, and partnerships leaders capable of negotiating protocol-level integrations with the AI tool chains their customers already run.