Nick Perry shows how to run a Level-10 meeting with AI agents by giving each one a distinct persona in its system prompt. Without deliberate design, AI agents agree or stay silent. Assign one agent as a growth strategist, another as a constraints manager, and a third as a research analyst to create the productive debate that drives better decisions.
So if I run a level 10 meeting with multiple AI agents in here, will they challenge each other's point of view and then we can come to consensus in like a level 10 structure meeting? Short answer, yes, but you have to architect it that way. By default, agents won't challenge each other. They'll just agree or stay quiet. You need to build that into their system prompts. Here's how to do it. One, Give each agent a distinct perspective. Fiverr, James, CEO, growth, revenue, big picture. Andrew, execution, constraints, feasibility. Scout, research, data, market signals, risks nobody's seeing too.
