What to practice as a UX designer when AI can generate the screens

Separate tool speed from product judgment

AI can help produce drafts and alternatives. A convincing output does not establish that the problem matters, the evidence is sound, or the proposed interaction works for the people who need it. Your practice should cover both making and evaluating. Treat generated screens as material to inspect, not an instruction to stop thinking.

Practice turning requests into researchable questions

When someone asks for a dashboard, ask who needs it, what decision they are trying to make, what they use today, and what goes wrong. Write a task statement and list the assumptions behind the proposed solution. This exercise improves your ability to connect interface choices with user needs. It also gives an AI tool a better brief if you later use one to explore alternatives.

Practice following evidence back to its source

Work with permitted research notes or a clearly fictional practice dataset. Separate direct observations from interpretations and recommendations. If AI suggests a theme, check the source material and record contradictory evidence. Do not let repeated model output become a substitute for participant evidence. Keep uncertainty visible in the artifact you share with the team.

Practice making and testing alternatives

Choose one small task and create two plausible approaches. State what each makes easier and what each makes harder. Try the flow yourself with realistic inputs, then observe relevant people where feasible and appropriate. A self-review is a useful start, but it is not user research. Record what surprised you and what you would change next instead of polishing the first output indefinitely.

Practice collaboration and explanation

Present the problem, evidence, alternatives, choice, and next test in five minutes. Ask a product or engineering colleague what constraint you missed. Then revise the decision brief. Your ability to make tradeoffs understandable helps other people contribute. A portfolio should show this reasoning alongside the finished design, with your role and the limits of the outcome made clear.

Try a four-week practice project

Week one: investigate one everyday task and write the problem brief. Week two: explore two flows and document the tradeoff. Week three: evaluate the prototype and record issues. Week four: revise and write a short case study. Use AI where it helps draft, vary, or critique, and note how you checked the result. This is a learning structure, not a certification or a promise of employment.

Choose support around the actual gap

If you need foundations, start with the free learning paths. If you need an experienced critique of a real flow, use the UX Decision Review. If your team wants a repeatable AI-assisted process, the AI for UX Workshop starts at $1,500 for one workflow, up to eight participants, a two-hour session, a template, and a review checklist. Tool subscriptions and custom integration are separate. Keep learning connected to work you can inspect and improve.

Keep exploring

A better-looking redesign is a hypothesis. Here is how to test it.

How to prioritize UX friction without giving your product a made-up score