Paper Number
ECIS2026-2136
Paper Type
SP
Abstract
We present an AI-assisted, semi-automatic approach for model transformation using platform-specific prompt templates and structured source-target mappings. The method transforms models created using domain-specific modelling language for health coaching chatbots into platform-specific representations for the Rasa Conversational AI framework by instructing a large language model. The approach is scalable to multiple target platforms, requires manageable effort, and involves a low learning curve. In this work, we focus on conversation models representing basic bot-user interactions. Our preliminary results demonstrate that the approach can generate functional, platform-specific artifacts from simple models, requiring only minor adjustments. The work will be continued to apply the approach to complex conversations involving technical abstractions and to extend it to additional conversational AI platforms.
Recommended Citation
Pande, Charuta and Hinkelmann, Knut, "AI-Assisted Model Transformation For Health Coaching Chatbots: From Domain-Specific Models To Multi-Platform Conversational Agents" (2026). ECIS 2026 Proceedings. 5.
https://aisel.aisnet.org/ecis2026/entmodel/entmodel/5
AI-Assisted Model Transformation For Health Coaching Chatbots: From Domain-Specific Models To Multi-Platform Conversational Agents
We present an AI-assisted, semi-automatic approach for model transformation using platform-specific prompt templates and structured source-target mappings. The method transforms models created using domain-specific modelling language for health coaching chatbots into platform-specific representations for the Rasa Conversational AI framework by instructing a large language model. The approach is scalable to multiple target platforms, requires manageable effort, and involves a low learning curve. In this work, we focus on conversation models representing basic bot-user interactions. Our preliminary results demonstrate that the approach can generate functional, platform-specific artifacts from simple models, requiring only minor adjustments. The work will be continued to apply the approach to complex conversations involving technical abstractions and to extend it to additional conversational AI platforms.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.