Keeping Community-in-the-Loop: A Novel Methodology for UX Research
M C, Clara Caldeira, Talita Pagani, Vinita Tibdewal · 2026 · Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA ’26) · doi:10.1145/3772363.3798828
Summary
This CHI 2026 Extended Abstracts poster from a Google research team introduces Community-in-the-Loop (CoLoop), a research and design methodology aimed at correcting the shortcomings of transactional, short-term UX research when working with people with disabilities (PwD) in the Global South. The authors argue that standard industry UX practice — designed largely in the Global North — relies on quick validation of internal hypotheses and therefore systematically misses the lived, intersectional realities of marginalized users, producing technology that is culturally distant and inaccessible. CoLoop is proposed as a structured five-day workshop format that braids together three established traditions: the velocity and cross-functional engagement of Google-style Design Sprints, the hands-on collaboration of Co-design, and the ethical depth and long-term partnership model of Community-Based Participatory Research (CBPR). It is explicitly anchored in the disability-rights principle of “Nothing About Us Without Us” and draws on Participatory Speculative Design and a feminist ethic of care. The workshop week runs Monday (team setup), Tuesday (Understand — counternarratives and technology primers), Wednesday (Experiment with high-stakes user journeys), Thursday (Ideate — speculative co-design of future technologies, closing with letters to the organization), and Friday (team debrief and analysis). Months of “decolonial learning” with local disability organizations precede the week, and community members participate in recruitment, facilitation, and co-analysis. The authors illustrate the method with two case studies: São Paulo (July 2024, 14 participants, reading technology and generative AI) and Bangalore (September 2025, 14 participants with visual, auditory, and cognitive disabilities, generative AI and fintech).
Key findings
CoLoop demonstrates that high-velocity product development and deep community engagement are not mutually exclusive when the engagement is structured around long-term local partnerships rather than one-off recruitment. Across the Brazil and India workshops, community members moved from passive consultants to active co-designers, producing concrete artifacts — a pipe-cleaner wheelchair asserting empowerment, a 3D paper card reading “I am not Disabled. Society is not Accessible,” and a speculative AI-companion robot — that reframed product team assumptions. Counternarrative exercises and “Superpower cards” helped participants reclaim their identities from dominant disability stereotypes before co-design began, which the authors credit for the depth of subsequent ideation. The São Paulo workshop surfaced how socioeconomic factors compound digital inaccessibility in Brazil and strengthened the local advocacy network by connecting previously siloed disability groups. The Bangalore workshop produced participant-designed agentic AI concepts that foregrounded deep personalization and multilingual context — dimensions typically absent from Global North AI roadmaps. A participant quote — “For the first time, we feel that our input is valued” — is offered as evidence of the method’s impact on participants themselves. Practical scaffolding (AI primers to reduce cognitive overwhelm, sensory decompression rooms, accommodations confirmed ahead of Day 1, facilitators who are themselves PwD) is reported as essential to the method’s psychological safety.
Relevance
For accessibility practitioners working inside large product organizations, CoLoop is a credible attempt to reconcile the tempo of industry UX with the ethical demands of disability-led research. Its novelty is less in any single component than in the packaging: a replicable week-long template that gives product teams permission to invest in long-term community partnerships without abandoning sprint-style delivery. The framework is directly useful for teams building AI products, procurement and research leads designing engagement with Global South users, and organizations trying to move past tokenistic co-design. Limitations are important to name: this is a poster-length contribution with only two case studies, both run by a single large company; success metrics are qualitative; and the “long-term partnership” claim rests on engagements measured in months, not years. The authors promise a full paper covering power dynamics and activity detail — worth watching for.
Tags: UX research · participatory design · co-design · community-based participatory research · disability rights · Global South · marginalized communities · generative AI · methodology
Standards referenced: UN Convention on the Rights of Persons with Disabilities (CRPD)