De-Sludging High-Friction Frontlines: Designing an AI-Mediated Service Gateway for Mental Health

Nearly one billion people suffer from mental disorders globally, yet 35 to 85 percent receive no treatment even when services are available, affordable, and known to them. This paradox suggests that the primary barrier to utilization is not only structural supply failure but also behavioral friction at the mental health service entry threshold. Drawing on sludge theory and administrative burden research, we conceptualize these barriers as service sludge—cognitive, emotional, and process frictions preventing vulnerable consumers from crossing the threshold. We adopt a design science research methodology and design MyMatcher, an AI-mediated service gateway—a pre-encounter frontline entity comprising three modules: a sensing engine that infers clinical intent from vernacular distress language, a low automated social presence (ASP) chatbot designed to reduce the social risk of disclosure, and a provider-matching module that collapses provider search into a curated shortlist. Through diagnostic and evaluative interviews with mental health practitioners, complemented by classification validation on a labeled corpus of over 4,000 texts, we derive three design principles and three research propositions for de-sludging high-friction service frontlines. Theoretically, we extend the 3A (Affordability, Access, and Awareness) framework by introducing Agency as a fourth behavioral dimension; we introduce the service gateway as a pre-encounter entity distinct from augmentation and substitution models; and we theorize an inversion of intimacy, the proposition that lower ASP facilitates sensitive disclosure in stigma-laden contexts, which qualifies a central assumption in service technology research. The framework is designed to generalize beyond mental health to high-stigma, high-complexity domains where behavioral frictions prevent consumers from acting on structural access to relevant services.