Agent retirement
operationsagent decommissioning · agent sunset · retiring AI agents
The decision and the work of switching off an agent that no longer earns its keep: who signs it off, what happens to its permissions, its data and the process it ran. This is an Implementa editorial formulation, not an established industry term.
Stated plainly so nobody misquotes us: this isn't a market category or a standard. It's what we call it, because we needed a name for the work nobody was doing. The context: discovering agents is already a product. Several security vendors sell inventory and detection of live agents. Deciding which ones are surplus and turning them off is still human work — and therefore still a service. Retiring an agent isn't deleting a deployment. It's answering five things: which process is orphaned and who picks it up, which credentials and access have to be revoked, what data and history are kept and on what legal basis, which integrations stop receiving its output, and what gets told to the team that relied on it. The metric we propose, and it's ours: agents retired over agents discovered. An inventory that only grows isn't governance, it's accumulation.
How it differs from
- Agent sprawl
- Sprawl is the problem: agents piling up unchecked. Retirement is the part of the cure you can't buy off the shelf.
- AI inventory
- The inventory tells you what you have. Retirement decides what you stop having, and carries out the consequences.
Examples
- A pilot summarisation agent with four months of no real usage and still holding read access to the CRM.
- Two agents from two different teams handling the same ticket type: one of them is surplus, and picking which is the decision.
FAQ
- When should an agent be retired?
- When any of three things is true: a full quarter with no meaningful use, it duplicates what another agent or person already does, or its running cost exceeds what it saves. None of the three is visible without measurement, which is why retirement starts with measuring.
- Why does the retirement ratio matter?
- Because it separates governing from hoarding. Showing how many agents you've deployed is easy; showing how many you've switched off and why is honest. Our rule of thumb: if you haven't retired one all year, you probably aren't measuring.