ai-user-research

What problem does it solve?

AI features often misalign with user expectations, causing confusion, misplaced trust, or avoidance; this Skill helps teams diagnose how users form mental models, when they trust or distrust outputs, and how errors affect behavior so you can design clearer, safer interactions.

Core Features & Use Cases

• Research objectives & framing: Define goals that map to product decisions such as adoption, safety signals, or feedback loops.
• Mental model elicitation: Protocols (think-aloud, analogy probes, card sorting) to surface user beliefs about how the AI works.
• Trust calibration & error recovery: Experiments to detect over-trust and under-trust, error message framing tests, and segmentation by expertise and stakes.
• Prototype testing: Wizard of Oz setups and usability workflows to validate outputs before model development, plus metrics (correction rate, re-generation rate, task completion) and feedback collection design.
• Use Case: Validate whether a content-summarization assistant is trusted for research tasks and design feedback UI that drives useful corrections.

Quick Start

Design an AI-specific user research plan for the new assistant feature that tests mental models, trust calibration, error perception, and feedback collection.