Vision
Vision
The Clerk, Not the Copilot
The AI industry's favorite metaphor right now is the copilot: your agent assists, learns, grows more capable, eventually becomes something like a peer. But talk to teams running agents in production and the vocabulary shifts entirely. They describe routing, filing, escalation, delegated authority. They are describing a clerk.
This matters more than it sounds. Agents don't gain power by getting smarter. They gain it by getting admitted into institutional systems that were never forced to articulate their own rules until something showed up that couldn't absorb them through hallway conversation.
The Clerk, Not the Copilot
The AI industry's favorite metaphor right now is the copilot: your agent assists, learns, grows more capable, eventually becomes something like a peer. But talk to teams running agents in production and the vocabulary shifts entirely. They describe routing, filing, escalation, delegated authority. They are describing a clerk.
This matters more than it sounds. Agents don't gain power by getting smarter. They gain it by getting admitted into institutional systems that were never forced to articulate their own rules until something showed up that couldn't absorb them through hallway conversation.
Evidence and Roles
The Clerk Pattern
A security agent compresses 3,456 telemetry rows into a structured alert. A coding agent produces a reviewable pull request. A payment agent executes transactions within cardholder-defined rules. Three domains, one convergent architecture: the agent assembles the case, the human acts on it. This is the conservative design surviving contact with real organizations. But a 680,000-chat experiment on Taobao suggests that knowing where the pattern fractures matters as much as understanding where it holds.
Review-Shaped Work
If agents handle the routine, what do the humans do? Not less. Something harder. The work reorganizes around policy authoring, exception handling, escalation design, and evaluating artifacts produced by a process you didn't participate in and may not fully understand. When that evaluation lacks context, authority, or the organizational backing to say no, you get governance theater: the signature is real, the scrutiny is decorative. The clerk model concentrates human responsibility into fewer, higher-stakes moments. Whether organizations actually redesign for that concentration is a separate and more difficult question.
Interface Futures
A chat transcript proves something was said. When the thing talking back is a person, that's often enough. But when an agent revokes access, issues a credit, or reclassifies a customs entry, a transcript is no longer adequate. What matters is what was done, under whose authority, and how you challenge it.
Agents are already moving fastest in domains where work was review-shaped before AI arrived. Pull requests, incident investigations, support cases, compliance filings. These aren't chat-friendly because the institution already knows how to argue with them. A security operations team reviewing an agent's evidence bundle evaluates activity timelines, severity scores, and remediation guidance. The output is a structured investigation object, not a conversation.
So the unit of interaction becomes less like a message and more like an action with standing: mandate, scope, evidence, consequence, and a path for appeal. Call it a docket. The approval button evolves into a bound technical object showing what's about to happen, what supports it, and what breaks if it's wrong.
A docket-shaped review catches procedural and factual failures more naturally than emotional ones. The file knows who called and what was offered. It does not know whether the person felt heard.
Further Reading
Past Articles
Much of what agent governance requires can be repurposed from existing compliance, QA, and security functions. But one r...
The EU AI Act's requirements for high-risk AI systems — logging, documentation, human oversight, accuracy monitoring — r...
When Visa published its agentic commerce rules this spring, the most striking thing about them was how familiar they loo...
An agent completes an enterprise workflow. The dashboard turns green. Three months later, someone asks who authorized th...