Chief Repeat Officer: Reminders and AI System Design

Jan 12, 2025 — Chief Repeat Officer

Executives are Chief Repeat Officers. Say the same priorities over and over until they stick.

This is just as true when your team includes AIs, but it looks very different.

With humans, you’re a repetition engine. You say goals and strategy in the all-hands. You say it in the 1:1. You say it reviewing the doc. You make a video saying it and share a doc saying it.

AIs need repetition even more than humans. Give the goals in the first prompt and it might be forgotten by message 10. The context window may be long, but attention drifts.

In one of my projects this weekend I have super tight persona and responsibility definitions and flows that ensure they are seen every session. I had documented that a Product Officer Agent should “own the positioning, packaging and persona - don’t worry about immediate next builds.” A few sessions later, that agent kept directing all its work towards the micro items on the roadmap.

AIs like tangible, concrete work in front of them and tend to forget meta things as they get buried in more tokens focused on detailed things.

So you have to build reminders into the system - not brute repetition but system design. Skills, loops, system prompts, context files—structures that keep surfacing what matters most. In this case I created a few hooks that bring back the primary directives within sessions to prevent drift.

You’re still the Chief Repeat Officer. You need to make sure the key things come back into your team’s context at runtime.

With humans, you repeat until it lands (at which point it’s probably about to change). With AI, you design the system so it repeats for you.

---