AI adds a cognitive layer to product work, speeding up reasoning and reshaping where designers focus. The job now moves from drawing interfaces to defining systems and behavior.
The traditional product design process was tidy and linear: Research → Synthesis → Wireframes → Prototypes → Handoff. That model made sense when interactions were predictable and screen-based.
Today, intelligence sits inside the product, interpreting, deciding, and acting. As a result, design is less about the UI and more about coordinating what should happen between human intent, machine behavior, and system boundaries.
This is how my workflow and design process have evolved while designing AI-infused products across identity and developer tooling.
Discovery: Faster, More Structured Synthesis
I still run interviews, gather raw notes, and immerse myself in context. The difference is how quickly I can move from noise to structure.
I feed transcripts into Dovetail or NotebookLM, then use Gemini or ChatGPT to surface themes, contradictions, and user goals. The value isn’t in outsourcing the thinking. It’s in shortening the path to clarity so I can focus on whether the problem is real, meaningful, and worth solving.
The guiding question shifts from “What did people say?” to “What should we care about?” This early phase used to take days or weeks. Now it often takes hours. Once the problem feels clear, I move directly into exploration.
Ideation: A Wider Possibility Space
AI is a powerful way to explore options. I start by asking it to propose alternate workflows, metaphors, or mental models. I’m not expecting the “right answer.” I’m looking for a range of options I would never have considered.
For early visual and conceptual exploration, I move between Mobbin (real-world references), Figma Make, Google Stitch, and Gemini. The goal isn’t to generate polished visuals. It’s simply to open up the space and invite surprise.
I now spend more time mapping behavior than drawing screens. If the system’s boundaries are unclear, interface polish doesn’t matter. Once behavior feels grounded, I return to UI. The interface is no longer the product. It becomes an explanation and control surface where users understand what happened, why, and what comes next.
UI becomes narrative, not destination.
Prototyping and Expression: Toward Shared Reality
To validate ideas quickly, I build functional prototypes with Figma MCP/Cursor or v0 by Vercel.
These tools help me simulate state, model logic, mock data, and pressure-test assumptions before involving engineering. I’m not trying to ship production code. I just want to feel how the system behaves, because behavior, not appearance, is what matters most.
With functional prototypes in hand, handoff becomes less about documentation and more about shared understanding.
What Actually Changes
AI doesn’t replace designers. It shifts where their judgment is most valuable.
Less:
- Pixel-level iteration
- Manual synthesis
- Speculative flows
More:
- Defining behavior
- Setting boundaries
- Communicating system intent
- Validating assumptions with working prototypes
The work becomes more architectural and strategic. More about reasoning than drawing. Designers move from screen-makers to system stewards who shape how intelligence behaves and how users stay in control. It’s a bigger job, and a more meaningful one.
Final Thoughts
Teams that treat AI as a collaborator, not just a feature, will explore more ideas, validate them sooner, and build products that feel coherent and trustworthy.
AI-first design isn’t about the tools themselves. It’s about adopting a new cognitive model, one focused on systems, behavior, and human judgment.
That’s the genuine transformation, and once you’ve experienced it, there’s no going back.