AI's Effect on Neurodivergent Knowledge Workers
A two-phase study with 27 participants combining qualitative interviews about Google Gemini on Workspace with a heuristic evaluation of 10 AI productivity tools, published as a white paper in partnership with Google and Blink, 2026.
The Challenge
As AI tools like Gemini became embedded in workplace environments, Google and Blink wanted to understand a critically underexplored question: how do neurodivergent knowledge workers, people with ADHD, autism, dyslexia, and related conditions, experience and use AI assistance at work?
Existing research on neurodiverse adults' use of workplace technology was narrow at best. The assumption that AI productivity tools worked uniformly across all users hadn't been examined, and the implications of that gap, for both product design and organizational policy, were significant.
Why This Matters
- arrow_right20% of the world's population identifies as neurodivergent
- arrow_rightMost neurodivergent workers choose not to disclose their status at work
- arrow_rightDesign for accessibility benefits neurotypical users too, the Curb Cut Effect
- arrow_rightAI features supporting executive functioning could be transformative for this population
Methodology
A two-phase approach that combined in-depth qualitative interviews with a comparative heuristic evaluation, providing both rich experiential accounts and structured cross-tool analysis.
Phase 1: Qualitative Interviews
27 one-on-one foundational qualitative interviews with a mix of internal Google employees and externally recruited participants who identify as neurodivergent (ADHD, autism, or dyslexia). Interviews explored daily workflows, AI tool usage, executive functioning challenges, and the experience of using Gemini on Workspace.
Phase 2: Heuristic Evaluation
Systematic heuristic evaluation of 10 AI productivity tools, including Gemini, Otter.ai, NotebookLM, and Goblin.tools, assessing how well each tool supported executive functioning needs identified in Phase 1. Developed evaluation criteria grounded in the interview findings.
Mix of internal Google employees and externally recruited participants. All identified as neurodivergent (ADHD, autism, or dyslexia). Mix of gender, ethnicity, age, and employment type. All were regular Workspace users with employer-paid accounts.
Key Learning
Researching underrepresented populations makes you a better researcher, full stop.
This project changed how I think about inclusive research practices in ways I didn't anticipate going in.
Working with neurodivergent participants required me to rethink some of my defaults. Putting interview questions on screen, for example, made a real difference for participants who struggle with auditory processing. It was a small adjustment that I now carry into other research contexts, because it turned out to benefit many participants beyond those who specifically needed it.
Participants also kept surprising me. Their relationship with AI was more nuanced than I expected coming in. Some described Gemini as genuinely transformative for their ability to do their best work. Others had developed deeply personal workarounds for the places where it fell short. Those stories shaped not just the findings but the way I framed the research questions as the project progressed.
One methodological detail I'm proud of: eight of the ten heuristics used in the tool audit came directly from themes surfaced in the qualitative interviews. Participants themselves told us what good AI support for executive functioning should look like. The audit was grounded in their words, not in a framework I brought from the outside. That connection between what people said in interviews and how we then evaluated the tools is something I want to replicate in future work.
Outcome
Published as a publicly available white paper, "The Future of Work: Google and Blink Examine AI's Effect on Neurodivergent Knowledge Workers," advancing Google's and the broader industry's understanding of AI accessibility and inclusive design. The research surfaced design recommendations that apply to AI product development well beyond the Workspace ecosystem.