Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstalliTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Uvi, the trading name of Naboo Research Ltd., raised a $7.5 million seed round on October 14, 2026, to develop a context layer that helps enterprise AI agents understand a company’s own systems, terminology and decisions. Uvi says that reducing repeated context-gathering can lower the cost of an agent answer from $0.75–$1.41 to $0.10–$0.16 in its benchmark—roughly an eightfold reduction, although the figures are company-reported rather than an independently audited industry measure.
What happened in Uvi’s funding round
CT Innovations led the seed round. Cardumen, Mana VC and 91 Ventures also participated. Uvi says it will use the funding to build a US team, expand sales and continue product development.
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL160_.jpg)
The company is an enterprise software platform, not a physical product. Its stated goal is to provide persistent business context underneath multiple AI agents and models instead of forcing each new agent to rediscover the same information.
Why enterprise agents keep rediscovering the same information
A newly deployed agent may be able to reason well in general but still lack the company-specific knowledge needed to act reliably. Business information is distributed across Slack, Jira, Confluence, Salesforce, code repositories, meeting notes and other systems. A question about a feature, for example, may require connecting a ticket to a pull request, a feature flag, a discussion thread and a later documentation change.
#1 Best Overall
Without those relationships, an agent often performs a search loop: it retrieves one record, searches for related terms, reads more records, and repeats the process before producing an answer. The extra retrieval and model calls consume tokens and add latency. Uvi co-founder and chief executive Gilad Salinger describes the cause as “structural, not about intelligence.”
What does Uvi do for AI agents?
Uvi describes its product as a context reasoning layer that combines a knowledge graph, entity mapping and a custom semantic layer. The service is intended to sit below different agents, so business context is not locked inside one model or one chat interface.
Connect related records in a knowledge graph
The graph links records that are usually separated by application boundaries. Uvi gives the example of connecting tickets, pull requests, feature flags, discussion threads and documents. An agent can then follow the relationships between those items instead of returning a disconnected list of search results.
Recommended Free Tools
Translate company-specific language
Organizations use internal abbreviations, product names and process terms that general-purpose models may not understand consistently. Uvi’s semantic layer is designed to map those terms to the entities and systems they refer to, allowing an agent to interpret a request in the company’s own vocabulary.
Reconstruct how decisions changed
Uvi also says it can reconstruct historical reasoning. A decision made in a chat thread may later be modified or reversed in a document comment; preserving that sequence helps an agent distinguish the current policy from an earlier conclusion. The company presents this historical linkage as part of context reconstruction, not simply document retrieval.
How can Uvi make enterprise agents eight times cheaper?
Uvi’s 2026 benchmark compares an agent using its context layer with a do-it-yourself agent searching the same enterprise sources. The company reports that the agents used the same model or an equivalent setup. Its reported cost figures are:
Rank #3
| Measure | Agent without Uvi | Agent with Uvi | Qualification |
|---|---|---|---|
| Cost per answer | $0.75–$1.41 | $0.10–$0.16 | Uvi’s company-reported 2026 benchmark; not an independent market average |
| Open-ended questions | Baseline not separately stated | Up to 46 times cheaper, according to Uvi | Reported for search-loop cases by Uvi and cited industry coverage |
| Response speed in open-ended cases | Baseline not separately stated | Uvi says answers were twice as fast | Company-reported comparison |
The ranges imply that savings vary by question and workload: comparing the endpoints can produce less or more than an exact eightfold difference. Uvi’s “eight times” description is therefore a summary of its benchmark, not a guaranteed reduction for every deployment. The reported advantage comes from avoiding repeated searches and model calls, rather than from changing the underlying model’s price.
The 46-times figure applies to open-ended search-loop cases, where an agent has to explore many possible sources. It should not be read as a general price for all enterprise questions. Uvi has not established an independent audit, standardized industry test or universal cost baseline for these numbers.
Security and permission enforcement
Uvi says it can run inside a customer environment or virtual private cloud. Its agent-aware permissions are enforced during graph retrieval: the graph should return only information that the person making the request is already authorized to see.
This design is intended to avoid adding a separate security retrofit after building a knowledge graph. In practice, buyers would still need to validate connector coverage, identity synchronization, revocation timing and audit requirements in their own environments; Uvi’s public description does not establish those implementation details.
Which models and agent interfaces can use Uvi?
Uvi says its context remains outside any single model and can be used with Claude, Gemini, GPT, Claude Code, Cursor and other agent interfaces. That portability is important for organizations that change models, use different tools for coding and operations, or want to test multiple providers without rebuilding their business-knowledge layer.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Uvi’s public materials describe broad compatibility, but they do not provide a complete connector matrix, supported-version list or integration service-level commitments. Those are questions to settle during technical evaluation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Customers and an example of context reconstruction
Independent reports identify monday.com, Lemonade and Melio as customers. Uvi’s own site also features testimonials from Melio and Global-e.
In the monday.com example, Jira tickets, pull requests and feature flags are linked so an agent can trace how a change moved from request to implementation and rollout. That relationship-aware answer is materially different from returning a set of individually relevant tickets with no explanation of how they connect.
What investors see in the market
CT Innovations investor Gwen Cheni said, “Most are still very far away,” referring to companies’ attempts to build context at scale. The comment reflects the gap Uvi is targeting: many organizations can connect a few data sources for a pilot, but maintaining an accurate, permission-aware context layer across systems and over time is harder.
The funding gives Uvi resources to expand its US presence and sales effort while continuing product development. The round announcement does not establish a valuation, revenue figure or profitability metric.
What companies should verify before adopting a context layer
- Answer economics: Reproduce cost per correct answer on your own workloads, including connector, storage and model fees, rather than assuming the reported eightfold reduction.
- Context quality: Test whether the system resolves internal names, links related records and distinguishes superseded decisions from current ones.
- Permissions: Confirm that retrieval follows the requesting user’s access rights across every connected source and that revoked access is reflected promptly.
- Portability: Check which models, coding tools and agent interfaces are supported in your intended architecture and how easily they can be changed.
- Coverage and operations: Request the connector list, synchronization behavior, failure handling, monitoring and audit exports needed for production use.
Bottom line
Uvi is betting that enterprise agents need a durable, shared representation of business context rather than another isolated retrieval workflow. Its $7.5 million seed round will fund US expansion and product work. The company reports answer costs of $0.10–$0.16 with its layer versus $0.75–$1.41 without it, plus up to 46-times savings in open-ended search cases, but those results remain vendor-reported. The key question for a buyer is whether Uvi can reproduce the same gains while preserving accurate history, model flexibility and user-level permissions across the company’s actual systems.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
