All field notes

AI as Your Design Partner: A Working Method for Designing Products With AI

Twenty years of product design taught me where AI helps, where it flattens everything, and how to tell the difference.

I've spent more than twenty years designing products — enterprise systems at places like NCR and Disney, and a lot of early-stage tools since. AI has changed my day-to-day work more in the last two years than anything else in that time.

But not in the way most people expect. AI didn't replace the design process. It made the cheap parts nearly free, which means the expensive parts — judgment, taste, and understanding people — matter more than ever. Here's the method I use to design with AI as a real collaborator.

1. Write the brief like you're briefing a senior designer

The most common reason AI design output looks generic is a generic brief. "Design a dashboard for my app" gets you every dashboard ever made, averaged together.

A good brief includes who the user is, what they're trying to do in this moment, what they're worried about, and what success looks like for them. Add constraints: your brand, your existing patterns, what this screen must not do. The quality of the brief sets the ceiling on the quality of the output.

2. Use AI to go wide, then you go narrow

This is where AI is genuinely transformative. Ask for ten structurally different approaches to the same screen — not ten color variations, ten different ideas about what the screen is for. A list-first layout. A single-question flow. A summary with drill-down. A conversation.

Then you do what AI can't: choose. Look at each through the eyes of the specific person from your brief and ask which one gets them to their goal with the least effort and the least anxiety. Usually you'll combine two.

3. Ask AI to critique, not just create

Some of the most useful design feedback I get now comes from asking a model to tear a screen apart. Try: "You're a first-time user who is short on time and slightly skeptical. Walk through this screen. What's confusing? What would you click first, and why? What would make you leave?"

Then ask it to review the same screen against accessibility basics: color contrast, text size, labels for screen readers, keyboard navigation. It won't catch everything, and it's no substitute for testing with people who use assistive technology, but it catches the obvious misses early, when they're cheap to fix.

4. Keep a design system the AI can follow

AI drifts. Ask it for five screens and you'll get five slightly different button styles. The fix is the same one design teams have always used: a design system. Write down your colors, type scale, spacing, components, and voice, and give that document to the AI every time.

The more specific the system, the more the output looks like your product instead of everyone's product. This is the single biggest difference between teams that get distinctive work from AI and teams that get templates.

5. Prototype fast, then put it in front of people

AI tools can turn a design into a working, clickable prototype in minutes. Use that speed for its best purpose: getting real reactions sooner. Five short sessions where you watch someone try to complete one task will teach you more than any amount of AI critique.

AI can help here too — drafting the test script, summarizing session notes, spotting patterns across sessions. But the moment someone pauses, frowns, and says "wait, where did it go?" is something you have to see for yourself.

6. Protect the parts that make it yours

AI averages. That's what it's built to do. Left alone, it pulls every product toward the same safe middle. Your job as the designer is to protect the specific choices that make your product feel like it was made by someone who understands its users: the tone of an error message, the one moment of delight, the thing you deliberately left out.

Where this leaves designers

If you're a designer, this is a moment of opportunity. Everything you've learned about people, systems, and judgment is now the scarce part of the process. AI handles more of the production; you get to spend more time on the decisions that matter.

If you're a founder without a designer, AI gets you further than ever — and a few hours with someone experienced can keep you from polishing the wrong flow. If you'd like a second set of eyes on your product's experience, the Growth Fix starts with exactly that: finding where people get stuck, and fixing the part that matters most.

Find where your users get stuck →

Want this applied to your business? The Growth Fix is one fixed-price session to find where your customers drop off and fix the part that matters most.

Ashley Kays

Ashley Kays

Founder

Founder of Waymaker. BigCo veteran (NCR, Walt Disney World, Wyndham Worldwide) turned solo operator. Building the operating layer above AI building tools.