AI doesn'tstart things.Humans do.
Every product begins with a human need, a business intent, and a feeling that no algorithm can generate on its own. My approach to AI in product design starts here — and keeps humans in the loop at every step after.
“The best use of AI is not to replace human judgment — it is to free humans to use it more.”
Four things I believe about AI in design
Before using any tool, I start with a clear philosophy. These four beliefs shape every decision I make about where AI helps — and where it shouldn't go alone.
Humans are the start point
AI cannot initiate. It needs human intent, context, and empathy to have anything meaningful to work with. Every process begins with a person understanding what is truly needed.
The loop is conversational, not one-way
AI should not just execute commands. It should question, flag gaps, and surface what it doesn't know — prompting humans to dig deeper rather than accepting a shallow brief.
Guardrails prevent hallucination
AI must know where to stop. Boundaries are set upfront — by humans — so that AI operates within a defined scope and asks for consent before moving beyond it.
Speed and quality are not a trade-off
When the human loop is maintained at every phase, AI eliminates busywork without cutting corners. Teams move faster on new requirements, and the quality of judgment stays high.
A conversational loop — at every phase
This is not a handoff model. At each phase of design, the human and AI stay in dialogue: human defines intent, AI questions and acts, AI flags what it's uncertain about, human decides what happens next.
human intent → ai questions → ai acts → ai flags gaps → human decides → loop continues
Interviews stakeholders. Understands the business intent, user need, and the feeling behind the ask — the things no dataset captures.
Surfaces patterns in what was shared. Asks: "Are these the right users? Is this the real problem? What's missing here?"
Points out conflicting business goals, underrepresented user groups, or unanswered questions before design moves forward.
Goes back to stakeholders. Digs deeper. Only a human can feel what is true versus what is just being said.
Guardrail —AI does not define the problem. It helps sharpen it. The human holds final say on the problem statement.
Speed and quality move together
When AI handles the busywork and humans handle the judgment, teams don't have to choose between moving fast and staying rigorous. Both improve — together.
Faster thinking, faster releases
AI frees teams from repetitive tasks — synthesis, conversion, variant generation — so human time goes toward thinking about what to build next, not reproducing what was already decided.
Quality stays because humans stay
Because human judgment is embedded at every phase — not just at the end — quality is not a final gate. It is a continuous thread. Nothing accumulates silently and ships wrong.
Shorter cycles, better iteration
AI compresses the time between idea and testable prototype. Teams get feedback sooner, learn faster, and don't release things that haven't been validated — just things that were built quicker.
Better business outcomes
Higher productivity contributes to revenue, retention, and team morale. The compounding effect: every release is a better starting point for the next one, because quality was maintained throughout.
“Move faster on what matters. Stay careful about what ships.”
This is how I work with AI.Not around it.
I see AI as a thinking partner that questions, synthesizes, and accelerates — within boundaries humans define. The result is design that is empathy-led, AI-amplified, and always human-approved.
— AJ · Product Designer