What would you need to trust an AI agent with real work?

Hey everyone, Nathan here from CopilotKit 👋

We’re getting ready to launch OpenBot, a self-hostable workspace for AI agents with built-in governance.

You can bring agents your team has already built through AG-UI, connect tools through MCP, and choose your models and infrastructure.

For business teams, that means control over how agents work:

• Grant tool access per Bot, role, and person.

• Require approval for sensitive actions and block others.

• Review an audit trail of what agents accessed and did.

• Run OpenBot on your own infrastructure.

Agent learning is part of this too. You can teach a process directly, or let agents learn from everyday work and turn successful approaches into reusable skills.

Ahead of launch, I’d love your perspective:

• What workflow would you hand over first to the bots you've created?

• What should always require approval?

• How would you want to review changes to an agent’s skills?

For a more thorough look, please check out: https://open.bot/