Abstract
Background: Frontline and emergency workers are frequently exposed to traumatic events, yet many report feeling underprepared or emotionally unsafe when engaging in conversations about trauma with peers. Stigma, limited access to training, and the lack of psychologically safe communication tools contribute to this challenge. To address the gap, the Trauma Conversation Simulator (TCS) was developed to support practice in trauma-informed dialogue through an interactive, AI-powered platform.
Objective: This study aimed to design, develop, and conduct an expert evaluation of a digital TCS prototype that enables users to practise engaging in supportive trauma-related conversations, with the goal of improving communication confidence, empathy, and psychological safety.
Methods: The TCS was developed using a co-creation approach grounded in the MRC-NIHR Framework for Complex Interventions. Built on LangChain using GPT-4o-mini, the simulator offers branching peer conversation scenarios with reflective feedback across five key communication phases. A panel of mental health professionals (n = 5) assessed the prototype using the Technology Acceptance Model (TAM), System Usability Scale (SUS), and qualitative feedback. Experts evaluated the tool’s usability, emotional realism, trauma sensitivity, and clinical relevance.
Results: Mental health experts identified strong alignment with trauma-informed principles and psychological safety. The AI-generated dialogue was described as emotionally attuned and realistic. Recommendations for improvement included greater nuance in the supervisor agent’s feedback, better visual differentiation across pathways, and enhanced conversational flow. Revisions were implemented to improve prompt quality and user interface clarity. The simulator demonstrated potential for use in healthcare, emergency services, and peer-support contexts.
Conclusions: The TCS shows early promise as a reflective, scalable training tool to improve trauma-informed communication in psychologically demanding roles. It fosters skill development in empathy, regulation, and peer support. The next stage will involve participatory co-design and testing with individuals with lived experience of workplace trauma, ensuring the tool reflects real-world emotional needs and supports inclusive, context-sensitive development.
Objective: This study aimed to design, develop, and conduct an expert evaluation of a digital TCS prototype that enables users to practise engaging in supportive trauma-related conversations, with the goal of improving communication confidence, empathy, and psychological safety.
Methods: The TCS was developed using a co-creation approach grounded in the MRC-NIHR Framework for Complex Interventions. Built on LangChain using GPT-4o-mini, the simulator offers branching peer conversation scenarios with reflective feedback across five key communication phases. A panel of mental health professionals (n = 5) assessed the prototype using the Technology Acceptance Model (TAM), System Usability Scale (SUS), and qualitative feedback. Experts evaluated the tool’s usability, emotional realism, trauma sensitivity, and clinical relevance.
Results: Mental health experts identified strong alignment with trauma-informed principles and psychological safety. The AI-generated dialogue was described as emotionally attuned and realistic. Recommendations for improvement included greater nuance in the supervisor agent’s feedback, better visual differentiation across pathways, and enhanced conversational flow. Revisions were implemented to improve prompt quality and user interface clarity. The simulator demonstrated potential for use in healthcare, emergency services, and peer-support contexts.
Conclusions: The TCS shows early promise as a reflective, scalable training tool to improve trauma-informed communication in psychologically demanding roles. It fosters skill development in empathy, regulation, and peer support. The next stage will involve participatory co-design and testing with individuals with lived experience of workplace trauma, ensuring the tool reflects real-world emotional needs and supports inclusive, context-sensitive development.
| Original language | English |
|---|---|
| Publication status | Published - 3 Oct 2025 |
| Event | Innovations in Psychology and Mental Health: The Next Generation - University of Strathclyde, Glasgow, United Kingdom Duration: 3 Oct 2025 → 3 Oct 2025 https://www.eventbrite.co.uk/e/innovations-in-psychology-and-mental-health-the-next-generation-tickets-1449280911219 |
Conference
| Conference | Innovations in Psychology and Mental Health |
|---|---|
| Country/Territory | United Kingdom |
| City | Glasgow |
| Period | 3/10/25 → 3/10/25 |
| Internet address |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 3 Good Health and Well-being
Keywords
- trauma-informed care
- psychological safety
- AI simulation
- frontline workers
- digital training
- co-creation
- lived experience
- conversational AI
Fingerprint
Dive into the research topics of 'Co-creation and expert evaluation of a trauma conversation simulator: developing an AI-based training tool to support trauma-informed peer dialogue'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver