Reflecti-Mate: A Conversational Agent for Adaptive Decision-Making Support Through System 1 and System 2 Thinking Conference: 34th ACM Conference on User Modeling, Adaptation and Personalization; June 08–11, 2026; Gothenburg, Sweden 34th ACM Conference on User Modeling, Adaptation and Personalization (UMAP ’26), June 08–11, 2026, Gothenburg, Sweden DOI: 10.1145/3774935.3806176 ISBN: 979-8-4007-2311-7/2026/06

Morita Tarvirdians, Senthil Chandrasegaran, Hayley Hung, Catholijn M. Jonker, Catharine Oertel · 2026

Read Reflecti-Mate: A Conversational Agent for Adaptive Decision-Making Support Through System 1 and System 2 Thinking Conference: 34th ACM Conference on User Modeling, Adaptation and Personalization; June 08–11, 2026; Gothenburg, Sweden 34th ACM Conference on User Modeling, Adaptation and Personalization (UMAP ’26), June 08–11, 2026, Gothenburg, Sweden DOI: 10.1145/3774935.3806176 ISBN: 979-8-4007-2311-7/2026/06 by Morita Tarvirdians, Senthil Chandrasegaran, Hayley Hung, Catholijn M. Jonker, Catharine Oertel (2026) 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-2605-22509.txt.

Introduction

Tarvirdians and team study an agent that helps with high-stakes personal choices by fitting the user's thinking style. Most support tools target thinking alone, not feelings and gut sense as well. The agent nudges users to blend thought types. In a study with 128 people, it led to broad, rich thinking, more personal thinking paths, and more joined-up thinking language. Users felt it gave stronger help for joined-up reflection. The plain tool, by contrast, gave samey sets full of thinking-only language.

Catalog record

CREATORMorita Tarvirdians, Senthil Chandrasegaran, Hayley Hung, Catholijn M. Jonker, Catharine Oertel
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46343
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.22509

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Reflecti-Mate: A Conversational Agent for Adaptive Decision-Making Support Through System 1 and System 2 Thinking Conference: 34th ACM Conference on User Modeling, Adaptation and Personalization; June 08–11, 2026; Gothenburg, Sweden 34th ACM Conference on User Modeling, Adaptation and Personalization (UMAP ’26), June 08–11, 2026, Gothenburg, Sweden DOI: 10.1145/3774935.3806176 ISBN: 979-8-4007-2311-7/2026/06 Morita Tarvirdians, Senthil Chandrasegaran, Hayley Hung, Catholijn M. Jonker, Catharine Oertel Abstract . Making high-stakes personal decisions involves cognitive, emotional, and intuitive processes, and individuals differ in how they allocate attention across these modes. Integration of these processes has shown to benefit decision making. Yet, most current decision-support systems focus primarily on supporting cognitive aspects, rather than adapting to the individual’s thinking profile to support integration of different types of thoughts. In this study, we investigate an agent designed to encourage integration by adapting to the individual user’s thought patterns. We explore its effects on participants’ perceptions of the agent and their reflective behavior, in comparison with unaided pre-reflection and a baseline agent. In a between-subjects study (N = 128), our agent, which fostered broad and elaborated thinking, enabled more personalized reflective trajectories, elicited more integrative reflective language, and was perceived as providing stronger support for holistic reflection. In contrast, the baseline agent produced homogenized profiles dominated by cognitive language across participants. † † footnotetext: Preprint. Accepted at the 34th ACM Conference on User Modeling, Adaptation and Personalization (UMAP ’26), 2026. 1. Introduction The process of decision making is an integral part of everyday life and involves a complex interplay between cognitive, emotional, and intuitive processes ( 49 ) . Each of these processes contributes distinct and complementary information, and prior work has shown that drawing on multiple sources of judgment is associated with more effective decision making, particularly in high-stakes contexts ( 15 ) . Such integration should take place during pre-decisional reflection ( 20 ; 16 ; 52 ) phase, when people think, elaborate and organize their thoughts before committing to a decision. Figure 1. ReflectiMate. A possible interaction with ReflectiMate where the user begins focused on internal aspects. Guided by its internal state, ReflectiMate gently probes overlooked experiential and external aspects through exploration and exploitation questions over multiple conversational turns. Based on the current state, the agent returns to internal aspects to probe underconsidered internal reasons, illustrating its aim of supporting integrative and elaborated reflection that adapts to the user’s dynamic thought patterns and implicit feedback. However, individuals differ substantially in how they draw on and integrate these processes. Some tend to rely predominantly on a single source of judgment while neglecting others ( 41 ; 42 ) . For example, when reflecting on a job offer that requires relocation, one person may focus heavily on anticipated excitement and fear of regret, repeatedly returning to emotional aspects while paying little attention to other considerations such as relocation logistics. Another person, in contrast, may engage in extensive cost-benefit analysis of salary and financial implications, while rarely considering emotional concerns such as distance from family or an intuitive sense of discomfort shaped by prior experiences. Despite the documented benefits of such integration, people often have difficulty to do so in practice. A key challenge is limited metacognitive awareness of one’s own thinking processes ( 43 ; 5 ) , which makes it difficult to recognize how different sources of judgment influence their decision. For instance, people frequently overestimate the extent to which their decisions are guided by cognition while underestimating the influence of intuition and emotion ( 35 ) , or fail to accurately identify which types of considerations dominate their reflection ( 49 ) . This lack of awareness makes it difficult for individuals to deliberately broaden or deepen their reflection in ways that integrate diverse considerations when making important decisions. Receiving personalized support is one potential way to help individuals overcome low-awareness cognitive patterns ( 14 ) . However, existing systems (e.g. ( 31 ; 39 ) ) often equate reflection support with promoting cognitive reasoning. Drawing on dual-process theory ( 26 ; 17 ; 48 ) , emotional and intuitive considerations are frequently treated as heuristic or bias-prone (System 1) and are therefore de-emphasized or left unsupported. As a result, adaptation in such systems is typically reactive, relying only on conversational history to generate follow-up questions that steer users toward further cognitive elaboration, regardless of individual reflection patterns. In this work, we address this gap by introducing a conversational reflection assistant that personalizes its dialog strategy to support integrative pre-decisional reflection (Fig. 1 ). Rather than aiming to alter users’ dominant reasoning mode ( i.e. if a person prefers cognitive reasoning we respect this rather than changing it to an emotional dominant reasoning style), our approach aims to support reflection by helping users surface potentially overlooked considerations and further elaborate underdeveloped thoughts. We present a user model that captures how individuals distribute attention across different thought categories and how their ideas evolve over the course of a reflective interaction. Building on this model, the agent dynamically balances exploratory questions that engage underrepresented types of thoughts with exploitative, Socratic follow-up questions ( 11 ; 37 ) that deepen existing ideas. This balance enables deliberate reflection—applying System 2-level elaboration to all thoughts, including emotional and intuitive ones typically associated with System 1—without privileging cognitive reasoning over other considerations. This implies that if a person indicates that not considering a certain aspect is deliberate, we respect the choice and do not try to change this. We evaluate our approach through a user study in the context of high-stakes life decisions, examining its effects on users’ reflective language use and perceived effectiveness of the support agent. We are answering the following research questions in this study: • RQ1: What differences emerge in reflective language during assisted reflection when participants interact with the experimental agent compared to a baseline agent? • RQ2: How do participants’ perceptions of the interaction differ between the experimental agent and the baseline agent? • RQ3: To what extent do the experimental and baseline agents preserve users’ dominant reasoning style? 2. Related Work Humans are prone to a range of well-documented biases and cognitive limitations during decision making process, such as confirmation bias, framing effects, and bounded rationality ( 53 ; 29 ; 6 ; 25 ) . These limitations can distort how individuals reason and form opinions. Therefore, a growing body of research has explored technological approaches to support people in the pre-decisional phase, when they form opinions that will guide their subsequent decision. For example, these approaches aim to help users overcome specific biases that can distort their reasoning ( 24 ; 46 ; 56 ) . Some studies have explored personalized interventions on information seeking behavior, as information seeking is an important aspect of the pre-decisional phase ( 16 ) . These systems often model how users interact with external information to provide adaptive support that facilitates less biased behavior. This could entail engagement with both sides (pro and con) of an argument ( 55 ) and various topics to disrupt self-imposed filter-bubbles ( 2 ; 3 ) . However, these studies do not support people’s own thinking process but rather focus on the process of exploring external information and opinions. In contrast, other

Read the complete text