Iron Choose the right AI employee
Search eligible agents by name, environment, scope, or trust before opening the conversation.
Agent ChatWork with your AI employees.
Work with your AI employees.
In one familiar conversation.
Open Chat, choose an eligible AI employee, and work through a normal text conversation. Follow activity while it works, review tool requests in the thread, open the resulting run, and give feedback on the reply.
How it works
A chat can stay familiar without hiding the governed work underneath.
Choose a deployed, chat-enabled agent, describe the job, and follow its working activity. If a tool action needs approval, the request appears in the thread with its destination, risk, trust, and approve-or-deny controls.
Product Walkthrough
Follow a conversation from agent selection to reviewed result.
Choose an AI employee, follow its work in the conversation, review a held action, and give feedback on the completed reply.
See work in progress
Keep the active turn, tool activity, cancellation control, and resulting reply in the same drawer.
Review what came back
Open the run behind a reply and add useful Good or Bad feedback without leaving the conversation.
Why it matters
What this changes once the work gets real.
Chat keeps the same controls
Policy, trust, approval, data protection, and allowed tools continue to apply inside the conversation.
Blocked content stays visible
A restricted message can remain in the thread with guidance to remove the sensitive value and try again.
Replies stay connected to runs
Open the agent run behind a completed reply when deeper operational detail is needed.
Make governed AI work feel like a conversation.
Keep selection, activity, approval, results, and feedback together in the Agent Chat workspace.