AI has fundamentally changed how I work
(MY PHILOSOPHY)
AI doesn't replace expertise.
It amplifies it.
Great software with AI
is built by engineers.
Great design with AI
is created by designers.
Great products with AI
are shaped by product teams
Below are just some of the ways I have integrated AI into my workflow when I see it as appropriate to use.
Discover
(HOW I APPROACH MY RESEARCH)
DEFINE
(HOW I REACH MY PROBLEM STATEMENTS)
Design
(HOW I IDEATE WITH AI)
Deliver
(HOW I DELIVER WITH AI)
I use AI to support the activity, but not to replace the human role within it.
(GREAT OUTCOMES STILL REQUIRE HUMAN JUDGMENT)
Conducting user interviews
AI can help me prepare questions, transcribe conversations and organize findings. But I prefer conducting moderated research myself. Some of the most valuable insights come from following unexpected threads, noticing hesitation, asking spontaneous follow-up questions and building enough trust for someone to explain what they actually think.
Facilitating workshops
AI is useful for developing workshop structures, exercises and visual material. Facilitation itself is fundamentally social. Reading the room, managing time, recognizing when a discussion needs to continue, involving quieter participants and resolving disagreement all require contextual judgment.
Making product decisions
AI can help structure research, surface trade-offs and challenge assumptions, but I don't delegate product decisions to it. Good decisions require teams to balance user needs, business objectives, technical constraints and organizational context. AI can bring information to the table. Accountability and judgment stay with the team.
Presenting research findings
AI can help structure a narrative and prepare supporting material, but I prefer presenting important findings personally. The discussion around research is often as valuable as the presentation itself. It gives stakeholders an opportunity to question findings, contribute context and understand the reasoning behind my conclusions.
Interpreting stakeholder intent
AI can process what was said, but understanding what someone means often requires organizational and interpersonal context. Priorities, competing interests, uncertainty and things left unsaid can all affect how I interpret stakeholder input. I treat AI's interpretation as another perspective, not as an authoritative reading of someone's intent.
Validating usability
AI can identify potential usability issues, inspect consistency and provide a useful first critique of a design. But it cannot tell me how the intended users will actually experience a product. When usability matters, I want evidence from representative users interacting with the product rather than treating an AI's prediction of user behavior as validation.
