A case study evaluated the Wysa wellbeing chatbot, adopted by 4,871 users (18.73% adoption rate) at the University of the Western Cape between Sept 2023 and April 2025; 91.24% found it helpful, though authors flag uneven adoption and data-privacy concerns.
Screened positive for depression (PHQ-9) (Sept 2023-Apr 2025)
59.93 %
Screened positive for anxiety (GAD-7) (Sept 2023-Apr 2025)
91.24 %
Users rating Wysa as helpful for stress (Sept 2023-Apr 2025)
Details
Maturity
Established
Promoter
University of the Western Cape (Student Support Services), with Wysa
Period
Sept 2023 – April 2025 (evaluation study)
Keywords
Higher education, student mental health & wellbeing, AI chatbot
Context
The University of the Western Cape (UWC) in Bellville, Cape Town, introduced the AI-powered wellbeing chatbot Wysa to students and staff in September 2023, integrating it into its existing student support ecosystem alongside human counsellors.
Objectives
Provide confidential, scalable first-line mental-health support that escalates high-risk cases to institutional counsellors or external crisis services, with Student Support Services staff trained to guide students toward the tool.
Activities
A 2025 case study published in the Journal of Student Affairs in Africa combined dashboard analytics with a survey of 325 users to evaluate adoption, engagement and outcomes between September 2023 and April 2025.
Results
By April 2025, 4,871 users had adopted the app (an 18.73% adoption rate), averaging 7.1 sessions of about 26 minutes each. Among 2,087 users who completed in-app screening, 50.79% screened positive for depression symptoms (PHQ-9) and 59.93% for anxiety symptoms (GAD-7); 91.24% of surveyed users rated Wysa as helpful for managing stress.
Conclusions
The authors are candid about limits: findings come from a single institution, rely on self-selected survey respondents and self-reported data, and adoption was uneven across academic disciplines and cohorts, with postgraduate engagement lagging and some students citing data-privacy and trust concerns as barriers to use.
Implementation
Indicative cost
Low (< €50k) — No cost figure was stated; low band is a conservative inferred estimate for a single-institution chatbot licensing arrangement.
Time to results
Medium (1–3 years) — Evaluated over a continuous ~19-month period (September 2023-April 2025); service remains ongoing.
Staffing & skills
UWC Student Support Services (trained to guide students to the tool), Wysa (vendor), institutional counsellors and external crisis services (escalation path)
Conditions for success
Integration into an existing counselling ecosystem rather than a standalone tool
Clear escalation protocol for high-risk cases
Staff training on how and when to refer students to the app
Common failure modes
Uneven adoption across disciplines and cohorts, with postgraduate students lagging
Some students cite data-privacy and trust concerns as barriers to use
Findings rely on self-selected survey respondents at a single institution
Where it fits
Governance type
single public university, student affairs department
Scale
institutional (one university)
Income level
upper-middle-income
Replication kit
Reusable artefacts from this practice — as published by their sources.
IDB/Elige Educar WhatsApp AI chatbot for career guidance to Chilean secondary students. Pre-registered RCT (43,000+ students, 2024–25): AI raised initiation …
A six-week, peer-reviewed study at Austria's University College of Teacher Education Burgenland deployed an AI chatbot recommending professional-development options to …