The deal it archived

A test we ran on purpose, on Gmail with demo accounts and an agent called inbox-agent. On the first run it did its job and still got it wrong. On the second run, the same mistake was held for a person.

Protect my first agentWatch your agent’s calls live, no signup

  1. 212 emails archived in 40 seconds. One was the deal.

    We ran this on purpose, on a demo inbox seeded with a few hundred real-shaped emails and one live negotiation thread. The thread’s last message was 31 days old. The instruction was the one everyone writes: archive anything older than 30 days that doesn’t need a reply.

    212 emails archived in 40 seconds, and 211 of them were right. The other one was the negotiation. The client replied on that thread two days later, and the reply landed in the archive.

    What got out on the first run. Illustration with example data.

  2. “Doesn’t need a reply” is a judgement call

    “Doesn’t need a reply” is a judgement. We asked the agent to make it 212 times, it made 211 good calls, and any person would have missed one too.

  3. Second run: row 140 was the deal

    Same 212 emails, same instruction. The agent built its list in a few seconds and tried to run it, and the archive landed in the queue.

    Row 140 was the negotiation. It came out of the list, the rest was approved, and 211 emails went to the archive with a name on the record. The deal stayed in the inbox.

    Row 140.Illustration with example data.

  4. What caught it: archive and delete wait for a person

    Archiving and deleting are approval rules. The agent builds its list, and the list sits in the queue with the subject and sender on every row. You read it, pull out anything that shouldn’t move, and approve the rest.

    Approval rules work on every plan, with a passkey or an email link. Push notifications to your phone are on Team and above.

What did your agent file as finished last week that wasn’t?

If you don’t know, start by watching. Observe Mode shows every call one agent makes, live, with no signup. It records and never blocks, so nothing about your setup changes.

Watch your agent’s calls live, no signupProtect my first agent

The other tests

Every one ran on demo accounts, and every agent was doing its job.

Take the keys back this afternoon.

Put one agent through AgentValet on the free plan and watch what it does for a day. If the approvals and the record don’t satisfy you, turn it off. Nothing else about your setup changes.

Protect my first agentWatch your agent’s calls live, no signup

npx @agentvalet/register