Startupdaily
startupTelegram

Your AI Agents Need a Chief Operating Officer, Too

AI·October 27, 2026

Your AI Agents Need a Chief Operating Officer, Too

Many companies have moved past asking whether to use AI agents and are now asking how to run them. In a guest column, Rick Folgmann, chief operating officer at Get Covered, argues that businesses should stop treating agents as just another software feature. His alternative is to manage them more like employees working in regulated roles.

That framing has practical consequences. Folgmann says each agent should operate from written rules, have its responsibilities spelled out, and report to a named human manager who answers for its performance. Without those pieces, it becomes hard to tell who owns a mistake or who is responsible for fixing it.

He recommends starting with tasks that already run on established rules, where the correct outcome is fairly well defined and errors are easy to catch. From there, the degree of autonomy should be tied to the risk of each workflow. A routine, low-stakes task can run with light supervision, while work with financial, customer or compliance consequences deserves tighter controls and more frequent human sign-off.

Expansion should be gradual. Teams should review agent output on a regular schedule and widen an agent's authority only after its results hold up over time. The column is aimed at organizations that have moved beyond pilot projects and need a repeatable way to govern agents as their use grows.

Reporting based on an external source.