Graphing the Everyday: A Neurosymbolic Approach to Eliciting Routines for Just-In-Time Adaptive Interventions
2026-08-10 • Human-Computer Interaction
Human-Computer Interaction
AI summaryⓘ
The authors studied how to turn people’s spoken stories about their daily routines into clear schedules using a combination of advanced language models and knowledge graphs. They found that these models struggle because human stories are often non-linear and complex, which doesn't fit well with the straightforward way computers process information. Also, the computer schedules often ignore how a person’s mood and energy change throughout the day. To fix this, the authors suggest design ideas like linking new tasks to current routines and making the system better at understanding users’ changing states. These suggestions aim to make digital health assistants more understanding and helpful over time.
Just-In-Time Adaptive InterventionsConversational agentsLarge Language ModelsNeo4j knowledge graphMental-model gapEcological mismatchEntity fragmentationRoutine piggybackingAdaptive negotiationSchedule tracking
Authors
Shakyani Jayasiriwardene, Blake Mountford, Meican Ma, Niels van Berkel, Nicholas Koemel, Matthew Ahmadi, Jorge Goncalves, Emmanuel Stamatakis, Zhanna Sarsenbayeva
Abstract
Just-In-Time Adaptive Interventions (JITAIs) increasingly rely on conversational agents to elicit user routines, yet translating fluid human dialogue into rigid schedule data remains a significant challenge. We conducted a qualitative investigation of a neurosymbolic pipeline, combining Large Language Models (LLMs) with a Neo4j knowledge graph, to map unstructured verbal narratives into actionable interventions. Through human-centric evaluation using natural-language playbacks, we identified a critical "mental-model gap," where the linear extraction of LLMs clashes with hierarchical, non-linear human storytelling, causing severe entity fragmentation. Furthermore, we articulate an "ecological mismatch," demonstrating that algorithmic schedule availability frequently ignores the user's fluctuating psychological receptivity and physical energy levels. To resolve these tensions, we propose actionable design heuristics, including routine piggybacking, adaptive negotiation, and scalable transparency. Ultimately, these guidelines provide a foundational framework for evolving rigid schedule-trackers into empathetic, context-aware proactive agents capable of supporting long-term health behavior change.