TETI, an LLM‑Augmented Telegram Chatbot for WHO Self‑Help Plus (SH+): Single‑arm mixed‑methods feasibility study protocol

Abstract Background : Workplaces expose employees to psychosocial risks that can undermine mental well‑being and contribute to stress and common mental health problems. Scalable digital conversational agents can reduce access barriers and provide brief, on-demand psychoeducation aligned with stepped‑care models. TETI will be a Telegram chatbot delivering WHO Self‑Help Plus (SH+) content, an ACT‑based program, augmented with large language model (LLM) modules to enhance personalisation and engagement. This protocol primarily aims to explore User Experience (UX) and Eser Engagement (UE) with TETI in a workplace setting, involving staff from the Bruno Kessler Foundation, Italy. Secondary aims are to assess acceptability and feasibility, and to explore proximal changes in distress, well-being, psychological flexibility, and knowledge retention. Methods : This mixed‑methods feasibility study will involve stakeholders in a workplace context, including domain experts in psychology, usability and potential end‑users, who will interact with the chatbot TETI on alternate days over two weeks. Participants will receive SH+‑consistent psychoeducational content (text, audio, visuals) via Telegram, with LLM‑supported conversational flexibility under predefined guardrails. The primary outcomes will focus on UX and UE, which will be assessed using a Semantic Differential scale, an adapted version of the uMARS, and the CES‑LCC. Qualitative post‑use semi‑structured interviews will be conducted to enrich the findings and provide further qualitative insights. Secondary outcomes will also be evaluated, including measures of distress and well-being, psychological flexibility, and a brief knowledge quiz assessing understanding of the intervention’s content. To determine the feasibility of the intervention, metrics such as recruitment, retention, interaction with the chatbot, and overall user satisfaction will be assessed. This study will involve assessments at three time points: baseline (T0), post-intervention (T1, after two weeks of chatbot use), and follow-up (T2, two weeks after T1). Discussion : This study will provide early evidence on the usability, acceptability and feasibility, along with early signs of change of a digital version of SH+, which will deliver psychoeducational content via a chatbot-based messaging system augmented with LLM support. Therefore, findings will help shed light on the strengths, limitations, and future directions for the digital adaptation of the WHO SH+ intervention, including the use of an LLM-augmented chatbot. Trial registration : In accordance with Italian Law No. 3/2018, art. 2, paragraphs 7 and 10, this study does not require ethics committee approval because: i) it focuses on the usability of a prototype digital intervention, ii) no patients or vulnerable populations will be involved, iii ) no clinical treatment or collection of sensitive health data is foreseen. Moreover, clear adherence to ethical principles is ensured, including voluntary informed consent and protection of participants' privacy.