MeetMap: Real-Time Collaborative Dialogue Mapping with LLMs in Online Meetings

TL;DR

MeetMap introduces real-time AI-assisted collaborative dialogue mapping for online meetings to address the nonlinear emergence of ideas and cognitive load. It offers two interaction variants, Human-Map and AI-Map, to balance user agency with AI scaffolding, and evaluates them against a business-as-usual baseline in a within-subject study with dyads. Results show MeetMap improves real-time tracking, shared understanding, and post-meeting coherence, with user preferences and behaviors differing by AI mode. The work provides actionable design guidance for integrating AI into synchronous collaboration tools to enhance sense-making without eroding human engagement.

Abstract

Video meeting platforms display conversations linearly through transcripts or summaries. However, ideas during a meeting do not emerge linearly. We leverage LLMs to create dialogue maps in real time to help people visually structure and connect ideas. Balancing the need to reduce the cognitive load on users during the conversation while giving them sufficient control when using AI, we explore two system variants that encompass different levels of AI assistance. In Human-Map, AI generates summaries of conversations as nodes, and users create dialogue maps with the nodes. In AI-Map, AI produces dialogue maps where users can make edits. We ran a within-subject experiment with ten pairs of users, comparing the two MeetMap variants and a baseline. Users preferred MeetMap over traditional methods for taking notes, which aligned better with their mental models of conversations. Users liked the ease of use for AI-Map due to the low effort demands and appreciated the hands-on opportunity in Human-Map for sense-making.

Figures (14)

• Figure 1: Users can collaboratively create dialogue maps in MeetMap. Users can use the AI-generated nodes to create the map. (1) Nodes are shown in Temporary Node Palette in real-time. (2) Users can drag nodes to the map and create links between the nodes. Using the interaction suite (a-d), users can create/delete nodes (a, d), edit nodes (b), and delete links (b). When creating/editing a node (b), users can specify the node content and node type.
• Figure 2: The visual representation of the IBIS notation schema: We use symbols to visualize the dialogue mapping notation schema, including "Questions, ideas, pros, and cons"
• Figure 3: We introduced a series of visualizations to improve the usability of MeetMap: (1) The discussion topics are segmented and labeled in the timeline; (2) When users click a topic in the timeline, the topic block changes colors, and the related nodes will be highlighted on the Map Canvas. (3) The timeline view represents speakers in different colors; (4) Mini-map for map navigation.
• Figure 4: Two Variants of the MeetMap System with different levels of human involvement and AI assistance. In Human-Map , AI only generates the nodes, and users will create the maps themselves. In AI-Map , AI generates drafts of dialogue maps, where users can further make edits in any way they want.
• Figure 5: AI-Map provides incremental generation of conversation maps so that users perceive the AI-generated content to be in real-time and digestible. (1) Summary nodes are generated in the Temporary Node Palette in real-time. (2) When a new topic is detected, the nodes in the previous topic are sent to generate a dialogue map. (3) As the conversation progresses, more small dialogue maps are generated.