Fitting the Message to the Moment: Designing Calendar-Aware Stress Messaging with Large Language Models Conference: Proceedings of the 7th ACM Conference on Conversational User Interfaces; July 8–10, 2025; Waterloo, ON, Canada Proceedings of the 7th ACM Conference on Conversational User Interfaces (
Pranav Rao, Maryam Taj, Alex Mariakakis, Joseph Jay Williams, Ananya Bhattacharjee · 2025
Read Fitting the Message to the Moment: Designing Calendar-Aware Stress Messaging with Large Language Models Conference: Proceedings of the 7th ACM Conference on Conversational User Interfaces; July 8–10, 2025; Waterloo, ON, Canada Proceedings of the 7th ACM Conference on Conversational User Interfaces ( by Pranav Rao, Maryam Taj, Alex Mariakakis, Joseph Jay Williams, Ananya Bhattacharjee (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-23997.txt.
Introduction
Pranav Rao and four co-authors test stress support messages written with each student's own calendar in mind, in a 2025 paper. Generic wellness tips often land at the wrong moment. Calendars quietly reveal which days will hurt.
Eight university students used a working probe app for one week. Each day it drafted a short message shaped by their upcoming events. Students liked messages that named their most stressful events first. They also preferred a brief, friendly tone over formal advice. The authors draw design lessons about asking the right questions. Tone calibration, they argue, decides whether such tools earn trust.
Catalog record
| CREATOR | Pranav Rao, Maryam Taj, Alex Mariakakis, Joseph Jay Williams, Ananya Bhattacharjee |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-23997
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Fitting the Message to the Moment: Designing Calendar-Aware Stress Messaging with Large Language Models Conference: Proceedings of the 7th ACM Conference on Conversational User Interfaces; July 8–10, 2025; Waterloo, ON, Canada Proceedings of the 7th ACM Conference on Conversational User Interfaces (CUI ’25), July 8–10, 2025, Waterloo, ON, Canada DOI: 10.1145/3719160.3737632 ISBN: 979-8-4007-1527-3/2025/07 CCS: Human-centered computing Empirical studies in HCI Pranav Rao, Maryam Taj, Alex Mariakakis, Joseph Jay Williams, Ananya Bhattacharjee Abstract . Existing stress-management tools fail to account for the timing and contextual specificity of students’ daily lives, often providing static or misaligned support. Digital calendars contain rich, personal indicators of upcoming responsibilities, yet this data is rarely leveraged for adaptive wellbeing interventions. In this short paper, we explore how large language models (LLMs) might use digital calendar data to deliver timely and personalized stress support. We conducted a one-week study with eight university students using a functional technology probe that generated daily stress-management messages based on participants’ calendar events. Through semi-structured interviews and thematic analysis, we found that participants valued interventions that prioritized stressful events and adopted a concise, but colloquial tone. These findings reveal key design implications for LLM-based stress-management tools, including the need for structured questioning and tone calibration to foster relevance and trust. 1. Introduction Stress is commonly defined as a psychological or emotional strain triggered by demanding circumstances ( Lazarus and Folkman 1984 ) . University students are particularly vulnerable, facing a combination of academic pressures, shifting responsibilities, financial concerns, and personal transitions ( Duffy et al. 2020 ; Barbayannis et al. 2022 ; Moore et al. 2021 ) . Prolonged stress in this population has been linked to serious psychological and physiological outcomes, including anxiety, depression, and cardiovascular disease ( Beiter et al. 2015 ; Osborne et al. 2020 ) . In response, numerous digital tools have been developed to support student wellbeing, offering strategies for relaxation, emotional regulation, and time management. While these interventions have shown promise ( Nahum-Shani et al. 2018 ; Kornfield et al. 2022 ) , a recurring limitation is their inability to adapt to the situational complexity of students’ lives, often leading users to perceive them as generic or misaligned with their actual needs ( Bhattacharjee et al. 2023a ) . Digital calendars offer a valuable but underutilized source of contextual information about students’ ongoing tasks and responsibilities ( Howe et al. 2022 ) . Many students use calendars extensively ( Mei 2016 ) , and the event details they input—titles, descriptions, and notes—capture meaningful, real-world pressures. However, this information is typically unstructured and difficult for traditional support tools to process. Advances in large language models (LLMs) offer a promising solution, as LLMs are capable of interpreting personalized and natural language content. Building on emerging research that uses LLMs to support wellbeing in educational settings ( Bhattacharjee et al. 2024c ; Kumar et al. 2024 ; Bhattacharjee et al. 2024b ; Rao et al. 2024 ) , we see an opportunity to connect LLMs and calendars to deliver adaptive, situation-specific stress support. Yet, this design space remains largely unexplored. Motivated by these opportunities and challenges, we investigate how students envision the role of LLMs in supporting stress management through their digital calendar events. Rather than designing full systems, we first aim to understand how students imagine LLMs could deliver more personalized and adaptive coping strategies. Specifically, we ask: • RQ: How do university students envision the role of LLMs in leveraging calendar data to tailor strategies for managing stress? To answer this research question, we conducted a one-week study with eight university students experiencing at least moderate levels of stress. Participants used a functional technology probe ( Hutchinson et al. 2003 ) that delivered daily text messages for stress support. Messages were adapted from existing content using an LLM to reflect participants’ digital calendar events. We found that participants expressed a desire to prioritize messages about stressful events and maintain a concise yet colloquial tone. This study contributes to human-computer interaction by examining early user expectations and responses to the integration of digital calendar data with LLM-generated, personalized stress management interventions. 2. Related Work Prior work has shown that personalization of stress-management interventions based on contextual factors can greatly increase their effectiveness ( Doherty et al. 2012 ; Paredes et al. 2014 ; Howe et al. 2022 ; Bhattacharjee et al. 2023b ; Nepal et al. 2024 ) . Timing is one such factor: just-in-time adaptive interventions, which aim to deliver support when users need it most, have been found to outperform randomly timed alternatives ( Nahum-Shani et al. 2018 ; Smyth and Heron 2016 ) . Several studies have explored using digital calendars to inform time-sensitive wellbeing interventions ( Howe et al. 2022 ; Kocielnik et al. 2013 ; Tateyama et al. 2022 ) . For instance, Howe et al. ( Howe et al. 2022 ) found that interventions aligned with Outlook calendar events increased engagement and reduced stress, while Kocielnik et al. ( Kocielnik et al. 2013 ) used calendar and sensor data to help users visualize and reflect on long-term stress patterns. Although this work highlights the value of digital calendars for identifying stress patterns and informing timing, it does not leverage the rich semantic content of the calendar events themselves. Our design introduces LLMs to interpret this content and support deeper personalization. By tailoring content to a user’s situation and emotional state, LLMs can make interventions feel more relevant and personally meaningful ( Kumar et al. 2024 ; Bhattacharjee et al. 2024b ; Bhattacharjee et al. 2024c ) . Recent studies have begun to explore the affordances of connecting LLMs with digital calendars for planning and scheduling ( Hoefer et al. 2025 ; Gundawar et al. 2024 ; Saxena et al. 2025 ) . For example, Hoefer et al. used LLMs to transform spoken narratives into structured calendar entries, allowing users to reconstruct their day with less manual effort ( Hoefer et al. 2025 ) . However, little is known about how such integrations might support stress management. We take an initial step towards addressing this gap by examining how students envision the role of LLMs in providing context-aware stress support using calendar data. 3. Design of the Probe We developed a lightweight interactive probe that delivered daily text messages tailored to the timing and content of participants’ Google Calendar events. The design of this probe was inspired by Rao et al. ( Rao et al. 2024 ) . Importantly, the probe was not designed to function as a fully adaptive intervention system, but rather to elicit participants’ expectations and reactions to this type of support. Each day, the system randomly selected an event from the participant’s calendar. Shortly before the event’s scheduled time, it prompted them to report how they were feeling and used GPT-4 to select and adapt a message from a curated set of expert-authored stress-reduction exercises ( Bhattacharjee et al. 2023a ; Meyerhoff et al. 2022 ) . GPT-4 received contextual information from the calendar and, when applicable, the participant’s check-in response. After each event, participants were asked to rate and briefly comment on the helpfulness of the message. A visual overview of the system’s message flow is shown in