In mid-2024, Fable’s VP of Innovation co-chaired a research group tackling accessibility testing with people who report challenges in memory, focus or learning. They drafted a screener to recruit participants who self-identify as having cognitive disabilities, reviewed existing studies, then ran a pilot with 25 testers. That pilot produced an interview guide and a survey to quantify usability experiences for people with cognitive needs. Next, the team teamed up with UC Irvine to compare findings from cognitive testers against a general-population group. They built three AI-generated prototype sites—a protein-recipe blog (Strong Snacks), a bookstore with filtering and swiping features (Turning Pages) and a hair-salon booking site (Crown & Comb). Thirty participants (five cognitive and five general on each site) completed task-based interviews and an Accessible Usability Scale (AUS) survey. Researchers logged every issue—confusion, missed elements, missing feedback—and every suggestion to improve content or layout. Cognitive participants uncovered 197 issues and made 93 suggestions; general-population testers found 113 issues and 54 suggestions. That’s roughly 1.8 times more feedback from the cognitive group. On Strong Snacks, for example, cognitive testers spotted an average of 9.8 issues versus 6.4 for others, and rated the site 13.7 points lower on the 0–100 AUS. Turning Pages, with more features and tasks, yielded similarly higher counts of problems and lower satisfaction from the cognitive cohort. These results back the idea that inclusive UX research uncovers more obstacles and richer design insights. In practical terms, screening for cognitive challenges and adapting interview practices revealed nearly double the usability gaps compared to typical user tests. Designers aiming for truly accessible experiences can’t skip this step.