Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Navigating the Unknown: A Chat-Based Collaborative Interface for Personalized Exploratory Tasks

Yingzhe Peng, Xiaoting Qin, Zhiyang Zhang, Jue Zhang, Qingwei Lin, Xu Yang, Dongmei Zhang, Saravan Rajmohan, Qi Zhang

TL;DR

CARE is introduced, a system designed to enhance personalization in exploratory tasks by combining a multi-agent LLM framework with a structured user interface, and was consistently preferred over a baseline LLM chatbot in a within-subject user study.

Abstract

The rise of large language models (LLMs) has revolutionized user interactions with knowledge-based systems, enabling chatbots to synthesize vast amounts of information and assist with complex, exploratory tasks. However, LLM-based chatbots often struggle to provide personalized support, particularly when users start with vague queries or lack sufficient contextual information. This paper introduces the Collaborative Assistant for Personalized Exploration (CARE), a system designed to enhance personalization in exploratory tasks by combining a multi-agent LLM framework with a structured user interface. CARE's interface consists of a Chat Panel, Solution Panel, and Needs Panel, enabling iterative query refinement and dynamic solution generation. The multi-agent framework collaborates to identify both explicit and implicit user needs, delivering tailored, actionable solutions. In a within-subject user study with 22 participants, CARE was consistently preferred over a baseline LLM chatbot, with users praising its ability to reduce cognitive load, inspire creativity, and provide more tailored solutions. Our findings highlight CARE's potential to transform LLM-based systems from passive information retrievers to proactive partners in personalized problem-solving and exploration.

Navigating the Unknown: A Chat-Based Collaborative Interface for Personalized Exploratory Tasks

TL;DR

CARE is introduced, a system designed to enhance personalization in exploratory tasks by combining a multi-agent LLM framework with a structured user interface, and was consistently preferred over a baseline LLM chatbot in a within-subject user study.

Abstract

The rise of large language models (LLMs) has revolutionized user interactions with knowledge-based systems, enabling chatbots to synthesize vast amounts of information and assist with complex, exploratory tasks. However, LLM-based chatbots often struggle to provide personalized support, particularly when users start with vague queries or lack sufficient contextual information. This paper introduces the Collaborative Assistant for Personalized Exploration (CARE), a system designed to enhance personalization in exploratory tasks by combining a multi-agent LLM framework with a structured user interface. CARE's interface consists of a Chat Panel, Solution Panel, and Needs Panel, enabling iterative query refinement and dynamic solution generation. The multi-agent framework collaborates to identify both explicit and implicit user needs, delivering tailored, actionable solutions. In a within-subject user study with 22 participants, CARE was consistently preferred over a baseline LLM chatbot, with users praising its ability to reduce cognitive load, inspire creativity, and provide more tailored solutions. Our findings highlight CARE's potential to transform LLM-based systems from passive information retrievers to proactive partners in personalized problem-solving and exploration.

Paper Structure

This paper contains 47 sections, 3 figures, 11 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Large Language Models for Exploratory Tasks
  4. Personalization and Inspiration in LLM-Based Systems
  5. User Interface Support for Thinking and Task Organization
  6. Motivating Insights from Empirical Studies on Chat-Based Interfaces
  7. CARE: Collaborative Assistant for Personalized Exploration
  8. User Interface
  9. LLM-Powered Multi-Agent Collaboration System
  10. Interaction Workflow among Agents
  11. Inquiry Agent
  12. Milestone Agent
  13. Needs Discovery Agent
  14. Ranking Agent
  15. Solution Craft Agent
  16. ...and 32 more sections

Figures (3)

  • Figure 1: Comparison of the UI and interaction styles between the CARE System and Baseline System. At the top is the CARE System, displaying the Conversation Panel, Solution Panel, and Needs Panel. The CARE System actively prompts the user, gathering their needs before creating a tailored plan. In contrast, the Baseline System, shown at the bottom right, features only a Chat Panel and tends to provide direct answers to the user's queries.
  • Figure 2: Overview of the CARE system. The gray area represents the User Interface, where users interact through the Chat, Solution, and Needs Panels. At the bottom, CARE's back-end consists of several agents, including the Inquiry Agent, Needs Discovery Agent, Solution Craft Agent, Milestone Agent, and Ranking Agent, which collaborate to process user inputs and generate personalized solutions. $\rightarrow$ represents user interactions, such as chatting or updating needs. $\rightarrow$ represents the internal data flow between agents. $\rightarrow$ represents that the agents write data to the interface. $\dashrightarrow$ represents that the agents retrieve data from the interface.
  • Figure 3: Comparative analysis of user responses to the CARE and baseline systems across five key aspects of user experience.