Designing for Complexity With Conversational Agent: FABER and Teacher Professional Learning

This article analyses FABER, a conversational agent designed to support teachers' instructional design from a reflective, inclusive, and sustainability-oriented perspective. Developed within the PNR – Progetto Azioni di ricerca per l'Istruzione e la Formazione and resulting from the collaboration between INDIRE and the University of Cassino and Southern Lazio, FABER integrates conversational AI, narrative design, a validated knowledge bank, and retrieval-augmented generation. Rather than acting as a prescriptive tutor, it functions as an epistemic partner that fosters professional reflection and the co-construction of meaning. The study adopts a qualitative, exploratory methodology based on co-design ateliers with teachers from different school levels. Teacher–chatbot dialogues were analysed through a hybrid framework combining thematic analysis, discourse analysis, and conversation analysis. Findings show that FABER supports the shift from operational requests to reflective design, promotes systems thinking, clarifies educational objectives, and fosters professional learning.