Designing accessibility guidance that appears exactly when developers need it
An AI reviewer that comments on pull requests — catching accessibility issues, explaining why they matter, and surfacing the right guidance inside the developer's workflow.
I recently built an AI-powered pull request reviewer for the CARE open-source project at Open Healthcare Network Foundation.
- Role
- Design + Engineering
- Timeline
- 2026
- Team
- Open Healthcare Network (CARE)
Read this case study as
Where the idea came from
The idea came from a pattern I noticed while working on accessibility issues. As accessibility improvements were being tracked through Jira tickets, I started thinking beyond fixing the current issues.
How can we help future contributors avoid introducing the same problems again?
The insight
In growing open-source projects, design systems, accessibility guidelines, and documentation are often evolving. Some knowledge is documented, some is still being defined, and some simply lives with the team.
That made me wonder:
What if contributors didn't have to search for the right guidance? What if it appeared exactly when they needed it?
What I built
So I built an AI reviewer that comments directly on pull requests. It identifies accessibility issues, explains why they matter, and surfaces relevant guidance within the developer's workflow—helping contributors learn while they're already coding.
🎥 Demo
Add a screenshot or short clip of the agent commenting on a pull request here.
Not just an AI project — a design project
For me, this wasn't just an AI project. It was a design project.
As designers, we often think about improving the experience of the people using the product.
But we can also design for the people building it. Sometimes, the best user experience starts with a better developer experience.