An LLM-Based System for Accessible and Personalized Scientific Communication

This paper presents the design and requirements engineering of A.L.E.S. (Augmented LLM-based Engagement System), an LLM-powered platform for accessible and personalized scientific communication. A.L.E.S. supports dual-mode interaction, scientific and simplified via a shared LLM engine and unified document corpus, addressing the needs of researchers, students, educators, and policymakers. The requirements analysis phase applied Agile and user-centered methodologies, supported by LLM-augmented simulation, to produce and prioritize functional, non-functional, architectural, and ethical requirements. The process produced 22 user stories across four epics, structured into a Requirements Traceability Matrix and aligned with stakeholder needs through hybrid prioritization (MoSCoW and AHP). Domain modelling incorporated stakeholder workshops, personas, and user journeys to ensure contextual relevance. The result is a validated and adaptable specification framework that ensures traceability, scalability, and compliance, positioning A.L.E.S. as a vigorous tool for enhancing the accessibility and societal impact of scientific knowledge.