Essence Coach: A Bot for Software Practice Adoption

Abstract

Although Essence has been proposed as a unifying framework for understanding and evaluating software engineering practices, its adoption remains challenging, calling for tools that can act as coaches for learners and practitioners. We present Essence Coach, a chatbot that integrates LLMs with retrieval-augmented generation (RAG) from a curated Essence knowledge base. Experiments with multiple LLMs show that RAG consistently improves response quality for Essence-related queries. Some preliminary tests suggest that such systems can help both technical and non-technical students understand and apply Essence concepts. While validation through broader user studies is needed, this work-in-progress highlights how LLMs can bridge abstract SE frameworks and practice adoption.

Shortened and revised version of the original preprint. Published in BoatSE '26