Allychemy: Leveraging AI to build a user-friendly accessibility checker

Web service teams are often asked to provide services that are difficult to do at scale — there can be challenges in access, budget and staffing resources. Accessibility compliance checking is a prime example. Tools and apps might be cost-prohibitive to provide on a recurring basis; online tools might require page by page checks. The UI of the tools are often difficult for users to understand and remediate errors they find. In a decentralized publishing envivonment where content contributors have varying skill levels, providing clear cut guidance is essential. Using AI, UF Health's web programming team used AI to help iterate and build an internal hosted application to deliver an easier to understand report than university-wide tools. Reports were sent to relevant parties with responsibilities to fix these WCAG violations, using plain language principles to convey the issue, why it needs to be repaired and how to do so. During this presentation, we will walk our audience through how we determined our project scope and used AI to iterate on the concept to speed up app development to provide the tool quickly to our content administrators across 800+ websites.