What would you trust an AI agent to negotiate on your behalf?

I’m building steer, an AI work assistant, launching here on Nov 22. Before then, I’d like to think through one design question with people who’d actually use it.

steer already captures meetings from your device, turns them into notes and to-dos, and follows up across your tools.

The new part is agent-to-agent negotiation: your agent talks to other people’s agents to find someone you need (a hire, a client, a mentor), checks the fit, and only brings back the matches worth your time. Nobody sees who you are until you approve.

The hard part isn’t the matching. It’s trust. So I’d like to hear from you:

• Would you let your agent chat with strangers’ agents for you? What’s off-limits?

• Do you want to okay every step, or just the final “sure, let’s meet”?

• If your agent could find you one person this month, who would it be?

I’ll reply to every answer, and I’ll share what we changed because of this thread.