Start by telling the AI what “non-judgmental” means to you, then give it clear boundaries and concrete examples to evaluate. The more specific your request and context, the more useful—and calmer—the feedback will feel.
Ask for “neutral, coaching-style observations” and explicitly ban labels and assumptions. For example: “No diagnosing, no moral judgments, no character statements—only behavior-based feedback and practical suggestions.”
Include what you did, how long, and what you were trying to build (e.g., “speaking up in meetings,” “social confidence,” “body language”). If the AI doesn’t know the target, it may critique the wrong thing.
Offer a short log: what happened, what you said, what you noticed in your body, and what you did next. If you have it, paste a script, journal entry, or role-play transcript. Concrete inputs lead to concrete feedback.
Request sections such as: “What worked,” “What to tweak,” “One small next step,” and “A 10-second reframe.” You can also ask for a 1–5 scale on specific skills (clarity, assertiveness, warmth) with brief justification.
Tell the AI how direct you want it to be: “gentle,” “balanced,” or “very direct but respectful.” If you’re feeling tender, ask it to prioritize encouragement first, then one improvement area.
Request two or three rewritten versions of what you said, plus a tiny drill you can repeat daily. That keeps feedback actionable instead of abstract.
For more examples and ready-to-use language, see the full guide: https://daryne.com/what-are-the-best-ways-to-ask-an-ai-for-specific-non-judgmental-feedback-on-my-confidence-exercises/.
Track one measurable behavior (e.g., number of times you spoke up, length of eye contact, or how quickly you recover after a stumble) for 1–2 weeks. If the trend improves and you feel more willing to try again, the feedback is likely helping.
Leave a comment