How to Present AI Work in Your UX Portfolio

Case studies are changing — here's how to document AI-assisted research, generative workflows, and human judgment so clients trust the process, not just the output.

Hiring managers and clients now ask how you used AI — not whether you did. A UX portfolio in 2026 should show where models accelerated work and where you made the calls only a designer should make. After shipping product work in Hong Kong, this is the case study structure I use when AI is in the toolchain.

Separate tool output from design decisions

Show the problem, constraints, and users first. Then explain which steps used AI (synthesis, copy drafts, image exploration) and what you edited, rejected, or validated with real users. Recruiters want judgment, not a screenshot of a prompt.

Document validation — AI does not replace it

Usability tests on final flows, not only AI-generated wireframes.
Accessibility and edge cases you fixed after automated suggestions.
Metrics or quotes that prove the shipped experience worked.

Be honest about limits and NDAs

If you cannot show proprietary models or data, describe the workflow in general terms: “used LLM-assisted thematic coding on interview transcripts” is enough. Never present raw AI output as final UI without noting your refinement.

The portfolio story is: AI sped me up; I stayed accountable for what shipped.

Optimize for humans and AI search

Clear headings, plain-language outcomes, and specific tools (Figma, ChatGPT, Claude, Perplexity for research) help both recruiters and AI search engines understand your expertise. Write for people first — structure helps citation in ChatGPT and Google AI Overviews as a side effect.

Let's work together

Open to UI/UX projects, collaborations, and product design support in Hong Kong and remotely.

Let's Connect