UX Collective: Designing how designers master AI

Mastery of traditional design software has a predictable shape: a new user learns the correct way to use Figma or Sketch, and over time their practice converges on techniques that more experienced practitioners also use. Amber Bouabdallah’s piece in UX Collective argues that AI tools break this model entirely, and that trying to teach AI mastery as if it were conventional software training produces the wrong outcomes.

Her core claim is that AI mastery is inherently personal and divergent rather than universal and convergent. When someone develops a sophisticated practice with an AI tool, it reflects their specific thinking style, domain knowledge, and workflow preferences — not a generalizable best practice that another designer should replicate. The implication for teams is significant: building AI capability inside an organization cannot rely on conventional training, because the goal is not to transfer a single correct technique but to give individuals permission and structure to develop their own.

Bouabdallah developed this thinking through a six-month peer learning series she co-designed with Ningdan Zhang at Salesforce, running monthly sessions across 46 designers. The format was deliberate: rather than expert instruction, sessions featured colleagues demonstrating their personal AI workflows. The emphasis was on showing real, incomplete practices, which created permission to experiment rather than a standard to measure against. Participants saw how AI practice looks when it is genuinely personal — different for each person, shaped by each person’s domain knowledge and habits.

One specific tension the piece names directly is the risk of losing exploratory design phases when AI tools encourage jumping straight to high-fidelity prototyping. The speed of generation can compress the divergent thinking that good design depends on, and Bouabdallah treats this as a legitimate concern rather than a limitation to be overcome with better prompts.

The article is most useful for design leads building AI training programs, UX operations practitioners deciding how to structure peer learning, and individual designers trying to understand why their AI practice looks different from their colleagues’ — and why that divergence might be intentional rather than a sign of doing it wrong.