Previous chapters used agents with predefined workflows — humans decided the pipeline, agents just filled the roles. An autonomous agent takes a high-level goal and figures out the steps itself.
This chapter builds a Business Idea Validator that autonomously:
Plans its own research questions
Executes each question with accumulated context
Synthesizes findings into a recommendation
Self-critiques with a confidence score
No human tells it what questions to ask — the LLM decides the entire workflow from a one-sentence goal.