Why Experimentation Is the Right Growth Mechanism: PRINCIPLE 20 - Grow by Experimenting with New Tools, New Roles, and New Ways

🔬 The Three Levels of Experimentation This principle operates at three levels. Progress at each level compounds into organizational capability. Level 1 — Tool Experimentation Actively try new AI tools, test their outputs against real work, and build genuine working knowledge of what they do well and where they fall short. Do not evaluate a tool based on a demonstration or a vendor overview. Evaluate it by using it on a real work item—something with actual acceptance criteria and a real consumer—and measuring what the output quality and cycle time look like in practice. Level 2 — Role Experimentation Experiment with the Planner, Producer, and Accepter functions across different work types, not just the ones most familiar or comfortable. A writer who has never taken the Accepter role does not know what acceptance looks like from the other side—and that gap is visible in the quality of their output. A manager who has never played Producer in an AI-augmented workflow does not understand the pace, the friction, or the places where the process breaks down. Role experimentation across the team builds the mutual understanding that makes collaboration genuinely effective. Level 3 — Ways-of-Working Experimentation Try a different daily cycle structure, a different handoff format, a different approach to acceptance criteria—and measure what changes. Each experiment, even an unsuccessful one, builds the organizational knowledge base that makes the next experiment faster and better-targeted. The goal is not to change everything constantly; it is to test assumptions about how work should flow and replace the ones that are wrong.