The Socratic method — using questions rather than answers to guide thinking — is old. It also happens to be the one coaching mode where AI genuinely outperforms most humans.
What the method is
Questions only. No answers. No advice. The coach never tells you what to do; they ask sharper and sharper questions until you arrive at the answer yourself.
Why AI is good at this
AI suffers from two failure modes that destroy ordinary coaching — sycophancy ("that's a great question!") and advice-giving. Both are fixable with system prompts. The Socratic mode bans praise and answers by design.
The three-question pattern
Question one: what are you actually trying to do? Question two: what would tell you you're succeeding? Question three: what is stopping you? Most coaching conversations end at question three with the answer you already knew but hadn't said out loud.
When to override the mode
Factual questions get factual answers. "What is the 2-minute rule?" should not get a Socratic deflection. The coach should switch modes based on the type of question. Good system prompts handle this; weak ones don't.
Further reading
Stanford Encyclopedia of Philosophy on Socrates is the academic primer. Comparing AI and human coaches covers why the Socratic mode particularly favours the AI.
About Sovereign — A privacy-first AI personal assistant that runs entirely on your iPhone. On-device LLM, zero-knowledge encryption, and a coach that learns from your own words. See how it works or visit the homepage.