Abstract
The high complexity of Digital Twins (DTs) creates barriers that restrict their accessibility to sustainability awareness functions. This research examines how Human-Centered Design (HCD) procedures together with Large Language Models (LLMs) enhance the usability while improving user engagement within digital twins. By integrating HCD principles and AI-driven interactions, the study proposes a framework to improve data interpretability and decision-making. The methodology was developed within the DISCOV.ER project (co-funded by the Emilia Romagna region and coordinated by CIRI ICT - University of Bologna), which tested sustainability interventions delivered through AI in digital twin systems. The study presents a structured method that combines AI technology with user-friendly digital twins and identifies future research paths as its final outcome.