Abhibhaw Asthana

Work / Devtron · AI agents

AI Control Plane: agents you can trust in production

Companies had no safe way to deploy AI agents or control what they do. I defined one control plane to build, deploy and govern agents.

Private betacurrent stage
3design partners shaping the product
User
Companies that want AI agents in production, and the platform teams that must keep them safe.
Problem
Teams could build an agent quickly. They had no safe way to deploy it or limit what it can do.
Decision
One control plane for the whole agent lifecycle: build, deploy and govern.
Not built
Another agent framework. Teams keep the framework they already use. The control plane governs how agents ship and run, not how they are written.
Status
In private beta with 3 design partners, not public yet. They shape what we build next. This page covers the problem and the product decisions only.

The problem

Building a demo agent got easy. Running one in production did not. Teams that shipped agents hit two questions:

  1. How do we deploy it safely? A new prompt or tool can break an agent in ways a unit test misses.
  2. What is the agent allowed to do? It calls real tools with real data.

Scripts and manual checks work for one agent. They fail at ten.

The decisions

  • One control plane to build, deploy and govern agents, so one team sees the whole agent.
  • Governance by default. A platform team sets the rules once. Agent teams ship inside them.

What I left out

Another agent framework. Teams keep the framework they already chose.