What happens when six AI agents with different perspectives discuss one decision instead of one chatbot giving one answer? I built a LangGraph group chat that lets a manager gather context, route the conversation, invite arguments and counterarguments, call a vote, and deliver a final verdict. Inspired by Edward de Bono’s Six Thinking Hats, the agents take the roles of pragmatist, skeptic, optimist, analyst, contrarian, and people person. This video follows the build from one agent to parallel responses, manager routing, a conversation loop, and debate mode.