Token costs are becoming the unglamorous problem at the centre of enterprise AI adoption. The more AI does, the more it costs to run, and the more people spend monitoring it — what Glean calls “botsitting.” The company’s counter is context: its Enterprise Graph indexes and permissions-manages an organisation’s knowledge once, so AI can retrieve it without re-searching or re-prompting on every query.
The benchmark Glean published at Glean:GO puts numbers on that claim. Compared against Claude Cowork on a suite of enterprise knowledge tasks, Glean reports 81% lower token costs ($0.58 per query versus $2.98) and preference in 78% of evaluated responses. Benchmarks produced by vendors about their own products are limited evidence, but the methodology — measuring real query cost against an enterprise knowledge baseline — is at least a testable framing.
The headline product announcement is Glean Tau, a desktop AI workspace built on an open-source harness. Tau unifies local files, applications, and code with Glean’s enterprise context layer, intended to handle multi-step tasks such as file organisation, document analysis, and cross-app coordination without constant human steering. Early use cases are weighted toward engineering workflows, though Glean positions it as applicable across enterprise roles.
On the governance side, Glean expanded its AI Gateway to cover more AI entry points, enforce restricted-topic policies via a component called Glean Protect, and extend governed MCP access beyond knowledge retrieval to include organisational skills and user memory. The expansion addresses a real pressure point: enterprises that adopted multiple AI tools through 2024 and 2025 are now trying to manage the resulting fragmentation.
Other capabilities announced include Glean Transform, which uses activity signals to identify where AI can absorb work from humans and measures the resulting impact; proactive task management that surfaces follow-ups and meeting preparation without a prompt; interactive dashboards combining structured data with unstructured business context; and team chat for shared AI workspaces.
Emrecan Dogan, chief product officer at Glean, said: “Models are getting better and more interchangeable. Understanding the enterprise is not. We designed Glean around that reality in 2019. The more AI moves from answering questions to actually doing work, the more valuable company context becomes, because it tells AI what matters, what should happen next, and how to act within the realities of the business.”
Glean, valued at $7.2 billion at its last funding round, is competing in a market that includes Microsoft 365 Copilot and Salesforce Agentforce for the enterprise AI platform category. Its core differentiation remains the argument that enterprise-specific context, rather than model capability, is the primary lever for AI performance in production environments.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
