Skip to content
How to ThinkIn the Age of AI

Question

What is cognitive agent design?

beginner

Quick Answer

New agents can inherit properties and patterns from existing successful agents rather than being built from scratch.

Example: A project manager has spent two years developing a reliable weekly review agent — every Sunday evening, she sits at the same desk, opens a specific template, reviews her calendar and task list, identifies the three most important outcomes for the coming week, and closes the laptop by 8 PM. The agent fires consistently. Its trigger is clear (Sunday 6 PM), its environment is controlled (home office, no notifications), and its output is specific (three written priorities). When she gets promoted and needs to manage a larger team, she needs a new agent: a monthly strategic review. Rather than designing this agent from nothing, she inherits from her weekly review. The monthly review uses the same physical environment (home office, notifications off), the same structural template (calendar scan, priority identification, written output), and the same time-bounded format (two hours maximum). What changes is the scope — monthly instead of weekly, strategic themes instead of tactical tasks, team outcomes instead of personal ones. The new agent does not need a fragile first thirty days of establishment. It inherits the environmental cues, the procedural scaffolding, and the self-discipline infrastructure that the weekly review already proved. Within two cycles, the monthly review is firing reliably — not because it was easy to build, but because it was built on components that were already working.

This concept is part of Agent Lifecycle in the How to Think curriculum.

Learn more in these lessons

agentsinheritancereusehabit-stackingtransfer-learningpattern-librariesagent-designbehavioral-scaffolding