01 / 08
AI is generating inaccessible code at scale and that's a design problem too.
Write accessibility requirements into your design specs. They're now AI training data.
LinkedIn — Carol Scott & Amy Liu
Bay Area Accessibility Camp · May 2026
02 / 08
The accessibility bug cycle is backwards. Catch it in design, not after launch.
Make accessibility criteria part of your definition of done, not a post-launch checklist.
LinkedIn — Carol Scott & Amy Liu
Bay Area Accessibility Camp · May 2026
03 / 08
MCP lets AI agents read your design system. Make it readable!
Annotate your design system for accessibility. AI reads what you write.
Deque — Jeremy Rivera
Bay Area Accessibility Camp · May 2026
04 / 08
Buried accessibility features don't help anyone.
Surface accessible features in your primary user flows, not buried in settings.
Bay Area Accessibility Camp · May 2026
05 / 08
"We have no disabled users" is a data collection failure, not a fact.
Recruit disabled users into your research panels. Absence of data is not absence of need.
Bay Area Accessibility Camp · May 2026
06 / 08
Context engineering is now a design skill.
Learn to write structured prompts and specs that give AI the right accessibility context.
eBay — Goutham Ponnada & Ian McBurnie
Bay Area Accessibility Camp · May 2026
07 / 08
Test with the right users before you commit to patterns.
Run assistive tech testing before locking in navigation patterns, not after.
Slack — Todd Kloots
Bay Area Accessibility Camp · May 2026
08 / 08
Treating accessibility as a one-time audit is a design culture problem.
Push for accessibility in every sprint review, not just before launch.
Asana — Devon Persing & Andrew Hedges
Bay Area Accessibility Camp · May 2026