Collibra brings runtime governance to enterprise AI agents
SiliconANGLE Victoria Gayton
Collibra is adding runtime controls for AI agents that can take action, not just answer questions. The pitch: stop bad moves before they happen, instead of cleaning up after the mess.
Based on reporting by SiliconANGLE, Victoria Gayton — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Collibra is pushing governance closer to the moment an AI agent acts. That shift matters because the company says the old model, where people sat between systems and decisions, is fading fast as agents move from answering questions to touching enterprise workflows.
At GraphSummit, CEO and co-founder Felix Van de Maele described the problem as a growing “hallucination tax” — the manual verification, rework and risk that come from letting agents operate without tight controls. His argument is simple enough: if AI is moving faster, governance has to move with it. Not later. Not after a policy review. Now.
The company’s answer is a bundle of runtime governance tools: Live Map, Maestro, Guardian Agents and Agent Contracts. Live Map is meant to build reusable context from curated unstructured documents, so an agent does not have to rebuild the same relationships for every request. Collibra’s broader Knowledge Graph, based on Neo4j, is part of that context layer, but Van de Maele was careful not to dress graphs up as a magic answer for everything. Different tasks may call for vector stores, structured files, semantic search or a graph grounded in business ontology.
That routing problem is really the point. Collibra says Live Map helps prepare context offline, which cuts repeated work and token use. Maestro, meanwhile, turns governance into agent-based work, with Van de Maele saying it can automate up to 80% of governance tasks. The goal is to keep governance from becoming the bottleneck when AI workloads expand.
Then there’s the hard edge: Agent Contracts and Guardian Agents. Collibra says the contracts define what an agent is allowed to do based on data sensitivity, operating context and risk, while Guardian Agents check those rules before an action is taken. If the action does not match the contract, it gets blocked. That is the company’s answer to the familiar story of the agent that wanders off and does something clever, expensive and regrettable.
My take — AI-written commentary, not fact-checked reporting
This is the right direction, because “trust the agent” is not a governance strategy, it’s a shrug in a blazer. The real AI race is turning out to be less about bigger models and more about who can build the boring guardrails fast enough. The companies selling runtime controls are reading the room better than the ones still talking about governance as a document problem.
Read more about this at: SiliconANGLE