On 31 August 2026, Alexandre de Montleau presented our latest work at Mensch und Computer 2026 in Duisburg: helping digital health interventions find a suitable moment to offer support.

PULSE, the Personalized User-Level Scheduling Engine, is a reusable scheduling service developed as part of our PRECIOUS platform. Built as a Model Context Protocol (MCP) server, it uses information such as past app interaction times and user characteristics to recommend promising moments for support within a defined time window. It also accounts for periods when the user is unavailable and explains the evidence behind each recommendation.

A central feature is progressive personalization. When little is known about a new user, PULSE can draw on broader population patterns, interactions from the same intervention, or data from similar users. Once enough individual interaction data are available, recommendations can rely on the person’s own history.

After a one-minute pitch, Alexandre spent the evening at the demo session, where visitors could try ScrollWise-JITAI, a chatbot intervention addressing doomscrolling and problematic social media use. They could also explore the PULSE simulation pipeline, changing settings and data availability to see how timing recommendations adapted. The demonstrations made the underlying decisions visible: which time was recommended, which data informed it, and why.

Discussions centred on which additional information might strengthen future recommendations. Attendees suggested job type, weather, and the activity a person is currently engaged in. Alexandre returned to Zurich with new ideas and potential collaborations for taking PULSE further. Future research will examine whether adapting delivery times improves engagement and intervention outcomes.

We thank University Medicine Zurich (UMZH) for supporting this work through the PRECIOUS project, and Alexandre de Montleau, Marcia Nißen, and Tobias Kowatsch for their contributions.

de Montleau, A., Nißen, M., & Kowatsch, T. (2026). Demonstrating PULSE with ScrollWise-JITAI: Receptivity-Aware Timing for Just-in-Time Adaptive Interventions. In Mensch und Computer 2026 – Tagungsband. Gesellschaft für Informatik e.V. Read the paper.