A playful but honest audit of the gap between what your AI can do and what your users actually trust it to do. This free interactive assessment scores your organization across 16 questions covering AI governance, trust design, and compliance, then generates a personalized risk report you can share with your team.
Four data points explain why AI governance matters right now: The EU AI Act imposes a €35M hard-cap fine for prohibited AI practices, enforceable across every member state. Only 35% of people globally trust AI companies (Edelman Trust Barometer, 2024). Zillow lost $304M when its AI pricing model failed with no feedback loop, a governance failure, not a technical one. And only 6% of organizations achieve significant EBIT impact from AI despite widespread adoption (McKinsey, 2025). The differentiator is not the models — it is operational execution: leadership oversight, workflow redesign, scaling, and governance.
Evaluates whether AI-related decisions have a clear, named decision-maker with documented authority. Assesses if organizational leadership treats AI oversight as a standing operational function, not a one-time initiative. Tests whether humans can meaningfully intervene during AI failure states; operators can override or recover from incorrect output without workflow disruption or data loss. Measures whether a structured mechanism exists to capture user-facing outcomes and route them back into model or product improvement.
Scores whether users have a reliable way to know what the AI will and won't do before committing to an interaction or decision. Evaluates whether the system communicates its reasoning, limitations, or confidence in a way users can act on. Tests whether users provide informed, meaningful consent before AI acts on their behalf. Assesses recovery quality after AI failure against defined standards. Monitors long-term user reliance patterns — over-reliance, automation bias, and disengagement. Measures whether a systematic mechanism exists to assess whether user trust aligns with actual system reliability.
Evaluates regulatory readiness including EU AI Act classification, NIST AI Risk Management Framework implementation, ISO 42001 alignment, and sector-specific requirements. Assesses data governance practices, bias monitoring capabilities, incident response procedures, and documentation standards. Tests whether engineering and product leadership can articulate the team's current AI interaction risk posture without preparation.
Each question is scored 0-3. Domain A questions are weighted 3x (max 36 points), Domain B questions are weighted 2x (max 36 points), and Domain C questions are weighted 1x (max 18 points). Total maximum score is 90. Risk tiers: High Risk (0-34), Moderate Risk (35-67), Low Risk (68-90).
This assessment is based on the AI UX design book "90% Human" by Konrad Piercey. The book covers the ESI framework (Educate, Simplify, Intrigue) for designing trustworthy AI systems, with practical strategies for AI explainability, trust calibration, and human-centered AI interaction design. Available on Amazon.
KNP Design Co. is a design consultancy specializing in AI UX design, AI governance, AI trust architecture, and AI compliance. Founded in New York City, with a decentralized global team in NYC, Berlin, London, and Shanghai. KNP helps organizations design, build, and govern responsible AI systems. Contact: awesome@knpdesign.co