Organization: PSI, Audere, Matchboxology and HSRC
Study objective:
- To assess the usability and acceptability of the coach chatbot by exploring user engagement, ease of navigation and overall satisfaction
- To explore user preferences and optimize the coach chatbot through iterative feedback
Intervention description: Audere, PSI, Matchboxology and HSRC are developing and evaluating an AI-driven Coach chatbot, with the aim of providing tailored, empathetic advice and support for people living with or at risk for HIV and TB, as well as those seeking broader sexual health information. Designed for scalability and accessibility, the chatbot aims to serve users who face practical, psychosocial, or interpersonal barriers to in-person care.
Inspired by South Africa’s high-performing Coach Mpilo peer-support model, which has achieved over 95% retention in care, the chatbot aims to replicate key informational and motivational functions of a real-life coach in a digital format. By using WhatsApp, the chatbot ensures discreet, user-friendly access to support anytime, anywhere.
While its primary focus is on HIV and TB, it is also able to address health issues such as STIs and overall sexual health and wellness. The Coach chatbot will undergo user testing through the Indlela B-Hub, focusing on usability, acceptability, and user experience, to refine the tool and generate insights for future digital health interventions targeting HIV, TB, and sexual health.
Evaluation description: This study will employ a mixed-methods design, using the implementation science Plan-Do-Study-Act (PDSA) cycle for user experience (UX) testing and refinement, to explore the usability and acceptability of the Coach chatbot in delivering HIV- and TB-related information and support. Data will be collected through post-interaction surveys and focus group discussions. The surveys will quantitatively assess users’ perceptions of the tool’s usability and acceptability, while the focus groups will qualitatively explore users’ experiences and engagement with the AI coach.
Study site: Participants will be enrolled in Indlela B-Hub sites in Gauteng and KwaZulu-Natal.
Target population:
Inclusion criteria:
- Voluntary self-reporting of HIV status (including those with unknown HIV status, HIV-negative status at last testing, and HIV-positive status)
- Comfortable consenting, talking and reading in English
- Currently residing in Gauteng or KwaZulu-Natal
- ≥18 years
Exclusion criteria:
- Not willing or able to travel to the sub-study sites
- Currently enrolled in another digital health or AI-related sub-study that may interfere with the outcomes of this research
- Not willing to consent to audio and screen- video recording
Sample size: The study aims to enrol 52 participants in round 1 and another 52 participants in round 2, across Gauteng and KwaZulu-Natal.
Duration: 6 months
