Stop building
a bot nobody
wants to learn

Six companies, one conversation, six times

Planning a week away means telling United, Delta, Airbnb, Marriott, VRBO and Wander the same things: who's coming, where you're going, what you'll eat, how you pay, which seat you want.

Every one of those agents is a stranger you have to brief. None of them talk to each other. All of them keep what you said.

Context does not survive the handoff

The obvious fix let the customer's agent brief each company's agent trades one problem for a worse one. Passing context agent-to-agent degrades it. Detail is dropped, intent is paraphrased, and by the third hop the receiving system is acting on a summary of a summary.

Delegation between agents is a lossy channel. An architecture that depends on it inherits the loss.

BYOA removes the hop. The customer's agent talks to your systems directly, against a surface you publish and control.

Own the environment, not the agent

A company-owned agent bundles four things that don't need to live together: what you know, what you can do, how you do it, and how you learn. Only the fifth the customer's history, preferences and standing context needs to sit with the customer. So put it there.

Their agent is a better support engineer than yours

Not because it's a better model. Because it was in the room for everything that led up to the problem.

The ticket: the speakers in the den cut out whenever the TV is on.

1. It set up this house's network last spring, so it already knows the SSID layout, which band the den is on, and that the mesh node behind the TV was added later.
2. It can reach the devices. It queries the speaker directly, pulls the firmware revision, and reads the actual channel it landed on.
3. It finds the receipt. A search of the customer's own mail turns up a range extender bought in March that nobody mentioned, broadcasting on the same channel.
4. It asks you the one question it can't answer itself whether your firmware has a known conflict on that band and you answer it from a published resource, with a version and a rule ID.
5. It writes the answer the way this customer likes to read it: short, dense, technical, no preamble.

Your agent could not have done step one, two, three or five. It would have opened with "have you tried restarting the device."

What you publish

Four layers, all machine-readable, all versioned. Serve them over MCP, an OpenAPI description, or both. None of it requires knowing which agent is on the other end.

{
  "outcome": {
    "state":       "blocked",
    "reason_code": "POLICY.RETURN_WINDOW_EXPIRED",
    "retryable":   false,
    "evidence": [
      { "resource": "policy/returns",
        "version":  "2026-08-01",
        "rule":     "RETURN-14" }
    ]
  }
}

You still get to improve

The usual objection: give up the agent and you give up the feedback loop. You don't.

You already observe every request that reaches you. Which tool was called, which schema failed, which policy version was cited, how many attempts it took, whether the task finished. That's a sharper signal than a transcript, and it arrives already categorized.

Aggregate it. When one in three attempts at a task class dies on the same missing parameter, that is a defect in your schema and you found it without anyone reading a customer's conversation.

The thing you improve is the surface. Every agent that touches it gets better at once, including the ones that haven't been built yet.