Uvi raises $7.5M to help AI agents learn business context and run at eight times lower cost

Artificial intelligence startup Uvi has raised $7.5 million in seed funding to address a costly issue for enterprises using autonomous AI agents. The funding was led by CT Innovations and joined by Cardumen, Mana VC, and 91 Ventures. AI agents, when deployed, know nothing about the company they're deployed to, requiring them to learn the company's business processes and operations from scratch.

This process costs significant computing resources and results in enterprises losing a fortune, according to Uvi's own benchmark tests. Uvi's solution is a "context reasoning layer" that provides a unified institutional memory for AI agents to tap into, instead of each agent having to relearn the company's processes. This layer is built atop a knowledge graph that functions as an infrastructure layer, sitting underneath every AI agent deployed by a company.

It uses entity mapping to understand the organization's engineering stack, track dependencies, and how different assets relate to each other. Uvi also analyzes historical context to reconstruct agent reasoning processes and has developed a custom semantic layer for the knowledge graph to help AI agents learn specific company jargon.

The cost savings are significant, with agents running on Uvi costing an average of just 10 to 16 cents per answer, compared to $0.75 to $1.41 without it. The company has already secured clients like Monday.com, Lemonade, and Melio, who benefit from Uvi's ability to understand the relationships between different tools and data sources, allowing agents to retrieve answers much faster.