Uvi Raises $7.5M to Give AI Agents the Company Context They Desperately Need

Uvi Secures $7.5M Seed Funding to Transform Enterprise AI Operations

Uvi, officially known as Naboo Research Ltd., has emerged from stealth with $7.5 million in seed funding led by CT Innovations and Cardumen1

1

2

2

. Mana VC, 91 Ventures, and Gaingles VC participated in the round, which the startup plans to use for building a US team, increasing sales, and continuing product development1

1

. The agentic AI startup is addressing a fundamental inefficiency that drains millions of dollars annually from enterprises deploying autonomous AI agents2

2

.

The Costly Problem of Lack of Contextual Understanding in AI

Each time a new AI agent is deployed, it begins with zero knowledge about the company that deployed it1

1

. These agents must learn everything about business processes and operations from scratch in a process that repeats again and again, racking up significant costs in computing resources2

2

. Company information is typically scattered across different tools—a decision might begin in a Slack thread, move to a Jira ticket, and finish in code1

1

. Until now, the only way agents could obtain this context was to reconstruct it from scratch with each new deployment, paying the model provider for each token used in that relearning process2

2

.

Eight Times Cost Reduction Through Context Reasoning Layer

Uvi's benchmark tests reveal that internally built AI agents cost companies approximately eight times more than necessary due to the need to repeatedly gather business context before starting work2

2

. The company estimated the cost per answer at $0.75 to $1.41 for traditional company-built agents, and in some cases exceeding $2, compared to $0.10 to $0.16 for agents using Uvi's solution1

1

2

2

. Beyond cost savings, Uvi's approach also cuts response times by more than half1

1

.

Building Unified Institutional Memory for Enterprises

Uvi has developed what CEO and cofounder Gilad Salinger calls a "context reasoning layer" that each company owns2

2

. Instead of having each new AI agent relearn how the business operates, they simply tap into this unified institutional memory made available to every AI tool within the IT environment2

2

. The software connects to Slack, Jira, Confluence, Salesforce, codebases, coding sessions, and meeting notes, allowing agents using models like Claude, Gemini, and GPT to understand what decisions were made and why1

1

2

2

.

Knowledge Graph Infrastructure Powers Agentic Workloads

The context reasoning layer is built atop a knowledge graph that functions as an infrastructure layer, sitting underneath every AI agent deployed by a company2

2

. It uses entity mapping to understand the organization's engineering stack, track dependencies, and understand how different assets relate to one another2

2

. Uvi also analyzes historical context, such as decisions made in Slack that were later reversed in document comments, to reconstruct agents' reasoning processes2

2

. The company has developed a custom semantic layer for the knowledge graph to help AI agents learn specific company jargon they'll encounter, ensuring accurate interpretation of terms2

2

.

Secure Access Controls Address Enterprise Concerns

Uvi's solution is secure by design, addressing the common challenge that knowledge graphs face with user permissions2

2

. The company built an access control model that operates at the application and administration layers, using data tagging to ensure employees only see information they already have access to1

1

2

2

. "We tag data so that a person who leaves a tool, or leaves the company, cannot be surfaced by agents down the road," Salinger explained2

2

. Each answer includes its original source, and the graph reconstructs relationships across platforms while carrying permissions with it1

1

2

2

.

Enterprise AI Adoption Drives Market Validation

Uvi is already being used by major enterprises including insurer Lemonade, work software firm monday.com, and payments company Melio, which accounting group Xero acquired for $2.5 billion last year1

1

. "Every enterprise we talk to has already spent a year and a real budget trying to build context-at-scale, and most are still very far away," said Gwen Cheni, who runs CT Innovations' Venture IQ Fund1

1

. Since the knowledge is stored outside any single model, companies can switch models without needing each new one to relearn the business1

1

.

Speed Emerges as Critical Bottleneck for Agentic AI

"Today, the conversation has shifted to cost and tokens. But as companies move into more complex agentic workloads, the real bottleneck becomes speed," said Gilad Salinger1

1

. He emphasized that in the early days of AI agents, most companies worried about accuracy and security, but now the conversation has turned toward cost, tokens, and speed2

2

. "The reason is structural, not about intelligence," Salinger noted, explaining that big companies run dozens of different tools with completely different data shapes, and the ability to pull multiple data slices in parallel and reason over that data doesn't exist for them2

2

. Even when an agent has completed a task, it would fail the same task the next time because it never retained how the work actually gets done2

2

.