I’ve always believed that great design comes from great systems. But for the longest time, those systems relied on people — catching the edge cases, flagging the inconsistencies, noticing what others missed. It worked, but it had limits. We could scale the team, but scaling how we improve — that was harder.
Lately though, something’s changed.
AI has quietly slipped into the design process — not as some magical design co-pilot, but as a practical partner. A tool that helps us design better, faster, and with fewer gaps.
The Real Value of AI Isn’t Flashy
Forget the headlines about AI replacing designers. That’s not the story I’m living.
What’s actually happening is far more useful:
AI is helping us catch things sooner.
AI is helping us speed up the boring parts.
AI is helping us think more systematically.
In my teams, we’ve started using tools like ChatGPT as a first-pass reviewer. It’s not perfect (and honestly, it shouldn’t be), but it’s pretty good at:
Running heuristic checks
Spotting consistency gaps
Reviewing flows against accessibility principles
Prompting designers to reflect on edge cases they may have missed
Here's a prompt we use to do a first pass and highlight small misses.
You are a seasoned UX reviewer and accessibility expert. I’m sharing a screenshot or mockup of a digital product. Please perform a comprehensive UX/UI audit of the design based on the following criteria:
1. Usability Heuristics (Nielsen’s 10 principles)
- Visibility of system status
- Match between system and the real world
- User control and freedom (undo, cancel, back)
- Consistency and standards
- Error prevention
- Recognition vs. recall
- Flexibility and efficiency of use (shortcuts, batch actions)
- Aesthetic and minimalist design
- Error recovery and guidance
- Help and documentation
2. UX Laws & Cognitive Psychology
- Hick’s Law (minimized decision-making)
- Fitts’s Law (tap target size)
- Law of Proximity & Similarity
- Miller’s Law (cognitive load)
- Gestalt principles (grouping, hierarchy)
- Tesler’s Law (hidden complexity)
- Von Restorff Effect (visual distinctiveness)
- Serial Position Effect (placement of key actions)
- Zeigarnik Effect (incomplete tasks/progress)
3. Design Fundamentals
- Typography: readability, hierarchy
- Color: contrast, consistency, accessibility
- Spacing & alignment: visual rhythm, cleanliness
- Visual hierarchy: clear distinction between primary/secondary actions
- Iconography: recognizability, tooltips
- Affordances: Do buttons/links clearly appear interactive?
4. Accessibility (based on WCAG 2.1)
- Color contrast (at least 4.5:1 for text)
- Keyboard navigation and focus states
- Screen reader compatibility (ARIA, landmarks, labels)
- Alt text and descriptions for images
- Readable font sizes and spacing
- Avoidance of motion triggers (if animated)
5. Content & Microcopy
- Clarity of labels, buttons, tooltips
- Tone and voice appropriate for the context
- Useful empty states
- Descriptive success and error messages
6. Flow & Navigation
- Overall structure: is it intuitive?
- Redundancy in access (e.g., search + navigation)
- Progress indicators or breadcrumbs
- Guidance for new users (onboarding)
7. Design System Compliance
- Are design tokens and spacing values aligned with the system?
- Are correct components and variants from the design system being used?
- Are layout patterns consistent with the documentation?
- Refer to: [design.innovaccer.com](https://design.innovaccer.com) for design system standards.
---
After evaluating, please do the following:
- Create a table of issues with severity levels (High / Medium / Low).
- Assign a score from 0–5 for each section (with rationale).
- Provide a final overall UX Score out of 5.
- Suggest the top 3 areas for improvements
It’s not perfect. But it sparks the right conversations — and surfaces things you may have otherwise missed. It’s like having a quiet design buddy who’s always ready with a checklist.
Not loud. Not intrusive. Just… there when you need them.
Why This Actually Matters
Here’s the thing: design quality improves when designers catch their own gaps. Not when someone else swoops in to fix them. When we give designers the tools to self-review — using smart, simple AI prompts — we’re not just improving files.
We’re improving designers. They start to think more systematically. They start to spot patterns. Their craft sharpens.
And as the team gets sharper, you don’t need to spend cycles chasing the small stuff.
You can spend time on the right problems:
Are our systems still scaling?
Can our components handle edge cases?
Are we supporting real user needs, or just designing the happy path?
The Real Shift: Less Policing, More Enabling
There was a time when I’d step into files to nudge things toward better quality. But that’s not where I add the most value now. Today, I care more about setting up the conditions where the system takes care of the small stuff.
Where designers grow faster because the feedback loops are tighter.
Where we use AI not to replace people, but to make our people better.
That’s how you scale design — not just by growing the team, but by growing the team’s ability to ship quality on their own.
A Quick Takeaway: Try This
If you’re thinking about where to start, here’s something simple we’ve found useful:
Set up a basic ChatGPT checklist for your design system. Things like:
Are tokens and spacings aligned?
Are we using the correct variants?
Are accessibility basics in place?
Are the empty/error/edge cases covered?
Encourage designers to run their work through it before it hits review.
It’s not perfect. It’s not fancy. But it builds better habits.
And it makes your review conversations so much richer.
Final Thought: AI Won’t Design For Us — But It Can Make Us Better Designers
I don’t think AI will take over design. But I do think the designers who know how to work with it will move faster, catch more, and think bigger. That’s the direction I’m leaning into:
Design smarter. Design faster. Design better — together.