HuePass

The Idea

While working on accessibility testing for a client project, I found myself diving deep into WCAG color contrast requirements. The existing tools felt dated, and I thought it would be useful to have a modern, streamlined tool that helps designers build color palettes with accessibility awareness baked in from the start.

At the same time, I was looking for a project to help me understand AI’s capabilities and limitations as a development partner. I wanted something contained enough to be a fair test, but meaningful enough to provide real insights.

HuePass became both: a genuinely useful accessibility tool and a laboratory for exploring AI-assisted development.

The Process

I approached this project the same way I work with developers on my teams. I would give clear tasks, let the AI complete them, conduct code reviews, request changes, and repeat. The rules I set for myself:

• No writing code unless absolutely required
• Limited guidance to knowledge I already had, no troubleshooting or research on the AI’s behalf
• Treat it like an MVP: focus on user experience and value delivery, not long-term architecture

The goal was to evaluate AI as a team member, not as a tool I was micromanaging.

What I Learned

The experience was eye-opening. The AI worked faster than any developer I’ve collaborated with and produced genuinely impressive results. It’s like working with a very capable developer who has encyclopedic knowledge of syntax and patterns.

But it has blind spots. Architecture decisions that seem obvious to an experienced engineer don’t come naturally. Performance implications of certain patterns aren’t considered unless explicitly raised. It’s excellent at the trees, less reliable at seeing the forest.

My takeaway: AI is still a ways off from replacing a senior architect for large, complex systems. The strategic thinking, the anticipation of scale challenges, the instinct for what will cause pain six months from now: that’s still firmly in human territory.

But I’m genuinely concerned for developers who want to focus solely on writing code without taking broader ownership of systems design. That middle ground is exactly where AI excels, and it’s only getting better.

The Result

HuePass provides real-time WCAG AA/AAA contrast checking, automatic color suggestions, color blindness simulation, and exports to CSS, SCSS, Tailwind, design tokens, and native mobile formats. It’s become a tool I actually use on client projects.

More importantly, it reshaped how I think about integrating AI into development workflows: not as a replacement for expertise, but as a force multiplier for those who have it.