The end-to-end control plane for AI has arrived

For years, AI governance meant reviewing models after the fact — auditing outputs, checking compliance boxes and hoping nothing went sideways between deployments. That era is over.

Today at the Data Innovation Summit in Stockholm, Collibra announced a new chapter in enterprise AI governance: the AI Command Center, a strategic partnership with Giskard, and momentum around our MCP Server that signals something bigger is happening across the industry.

AI isn’t just generating outputs anymore — it’s taking action

The shift to agentic AI is real, and it’s happening fast. Organizations aren’t just running models in isolation — they’re embedding agents directly into workflows, systems and decisions. These agents trigger processes. They influence real business outcomes. They operate continuously, often without a human in the loop.

Traditional governance frameworks weren’t built for this. They were designed for a world where AI was a tool you called on demand, not a participant woven into how your business runs. When agents are active in production around the clock, retrospective review is simply too slow.

What organizations need now is continuous context and control — not just governance as a checkpoint, but governance as an operating model.

Introducing the AI Command Center

The AI Command Center is a unified control plane that gives organizations real-time visibility and active management over AI systems and agents in production. It brings together agents, models and use cases into a single view — with live signals on ownership, activity, decisions and risk.

• See what agents are actually doing — not just what exists on paper, but what’s running in production right now.
• Understand decisions and their impact in real time — trace how agents are reasoning and what they’re changing downstream.
• Continuously monitor behavior and emerging risks — move from periodic audits to always-on observability.
• Intervene instantly when agents deviate — act before deviation becomes exposure.

More than 40 leading enterprises have already joined the private preview — a strong signal that the market has been waiting for exactly this kind of centralized AI operations capability.

Validation through partnership: Giskard and AI UC-1

Visibility into agents is necessary, but it’s not sufficient. Before deploying any AI system at scale, organizations need confidence that it actually behaves as intended.

That’s why we’re announcing a technology partnership with Giskard, which connects execution-level testing and validation directly to the Command Center. Model behavior, agent performance, and potential risks are continuously captured and fed back into governance — closing the loop between testing and production.

Alongside this partnership, Collibra is introducing out-of-the-box assessment templates aligned with the AI UC-1 standard — providing organizations with a consistent, repeatable way to evaluate AI systems across risk, compliance and readiness dimensions.

Together, these capabilities mean AI systems aren’t just built and deployed — they’re validated, assessed and continuously governed throughout their entire lifecycle.

Context is the new infrastructure: MCP adoption accelerates

As agent networks grow more complex, the quality of context they operate on becomes a critical variable. Agents that can access accurate, governed metadata and business context make better decisions. Agents flying blind make costly ones.

Collibra’s MCP Server enables agents to access governed metadata in real time — structured, trusted context that flows directly into AI workflows. The adoption numbers speak to real demand: more than 100 customers are now using the MCP Server in production, and it has reached a top position in the Databricks Marketplace.

We’ll be showcasing how MCP is reshaping enterprise AI deployment at the Databricks Data + AI Summit, where Collibra joins as a launch partner in the Agent Bricks ecosystem.

What this means for your organization

The companies that win with AI won’t just be the ones that move fastest — they’ll be the ones that move with the most control. Speed without governance creates risk; governance without real-time visibility creates blind spots. The AI Command Center is built to give organizations both.

If you’re deploying agents into production workflows today, or planning to, the question isn’t whether you need a control plane. The question is whether you have one yet.

Interested in joining the AI Command Center private preview or learning more about Collibra’s MCP Server? Reach out to your Collibra account team or visit collibra.com. For media inquiries, contact [email protected].

• 

  Teresa Rojas

  Teresa Rojas

  Senior Director of Product Management, AI Governance

  Collibra