Reflection Pattern: AI Agents Self-Correct in Production

Reflection pattern for LLMs: a self-critique loop that improves correctness, clarity, and reliability with measurable quality gates, ROI, and safety controls