Uvi has raised $7.5M in seed funding to fix a cost problem that grows every time an enterprise deploys another AI agent.
CT Innovations led the round, with Cardumen, Mana VC and 91 Ventures participating.
The inefficiency is structural. A newly deployed agent knows nothing about the company that runs it, so it must reconstruct business context from the ground up. Companies pay model providers for every token spent on that relearning, and the process repeats with each new agent. Uvi, formally Naboo Research, cites its own benchmarks showing internally built agents cost the average company around eight times what they should because of it.
Founder Salinger argues the problem is structural rather than a matter of model intelligence. Large organisations run dozens of tools with entirely different data shapes, and the ability to pull two slices of data in parallel from, say, Snowflake and GitHub and then reason across them does not exist off the shelf.
Uvi’s product is what it calls a context reasoning layer that each company owns. Instead of every agent learning the business anew, agents tap a unified institutional memory available across the organisation’s tools.
As agent deployments multiply, the argument is that context, not model capability, becomes the real enterprise bottleneck.
