Literary interpretation mediated by an intelligent conversational agent: A prompt-based framework for digital transformation in the humanities

Abstract

The digital transformation of higher education is reshaping literary learning, where interpretation requires sustained pedagogical support rather than ready-made answers. While large language models can provide personalised assistance, use without instructional framing may encourage surface-level engagement. This study introduces RELIA: empathetic reading guide with artificial intelligence, a conversational agent developed through an iterative design-and-evaluation approach to support interpretative learning. RELIA     uses structured prompts as an instructional script in a three-mode design—Tutor (guided support), Checkpoint (adaptive formative assessment), and Evaluator (formative learning summary), including a conditional review loop that redirects learners to additional support when repeated difficulty is detected. An exploratory mixed-methods study (N = 20) examined learner perceptions, complemented by thematic analysis of open-ended responses. Results provide preliminary indications of high perceived clarity (M = 4.13) and strong acceptance of adaptive formative assessment (M = 4.21), alongside high usability but more variable motivation (M = 3.83). Moving beyond an evaluation focused solely on RELIA, the article proposes a transferable framework for integrating conversational agents into humanities teaching–learning cycles, showing how structured prompting can mediate literary interpretation while preserving critical agency.