AX Framework

Design AI decision systems, by Marta Fernandez

AI changed the product. Design has to change with it. AX Framework is a practical framework for designing AI systems that understand, decide, act and evolve while keeping humans in control.

It includes seven pillars, the AX Working Model, canvases, Decision Maps, governance checks and evaluation toolkits.

The Seven Pillars

  1. Human-Context, decisions grounded in people, tasks and environments.
  2. Explain-and-Feedback, make decisions legible; invite corrections and signal uncertainty.
  3. Override & Fallback, clear exits, controls and human escalation at any step.
  4. Metrics / Co-Evolution, trust, adoption, recovery and fairness over time.
  5. Ethics & Privacy, consent, boundaries, least privilege by design.
  6. Adaptivity, fit to user goals and context; safe personalisation.
  7. Collaboration, shared rituals, artifacts and decisions across PO, UX, Eng and AI ops.

Mindset

  • AI Trust Mindset, designing for reliability and predictability.
  • Applied Ethics, ethical trade-offs and accountability in AI.
  • Human Experience, designing AI experiences around people.

AXEL

AXEL stands for AI eXperience, Ethics, Limits. It is the AX Framework materialised as a governance facilitator for AI products. AXEL does not answer for you, it makes you think.

Explore

Machine-readable summaries for AI assistants: /llms.txt and /llms-full.txt.