Are We On Track? AI-Assisted Active and Passive Goal Reflection During Meetings

Research Background and Issues

• What problems or challenges did the authors identify?
  Meetings often lack clear objectives and direction, leading to inefficiency. For instance, goals may become ambiguous during meetings, or discussions may go off-topic. Moreover, most meeting technologies focus on content dissemination but neglect support for meeting goals and reflection—a process of gaining new insights through reviewing and understanding experiences.

• Why is this issue important?
  Meetings are a critical component of teamwork and decision-making, but their inefficiency wastes resources and time. Introducing artificial intelligence (AI) to support meeting reflection can significantly enhance goal clarity and meeting effectiveness.

• Research Motivation and Related Work
  Although prior studies have explored meeting design, real-time feedback, and behavior prediction tools, these efforts largely remain superficial, failing to delve into how to support real-time goal reflection. Additionally, existing systems are often tested in low ecological validity environments, making it difficult to fully address the needs and challenges of real-world work settings.

Solutions

• What methods or solutions did the authors propose?
  The authors proposed two AI-supported reflection systems based on technology probes:

  1. Passive (Ambient Visualization, Viz): Provides continuous but low-disruption feedback by visually displaying the meeting's goals and topics in real time.
  2. Active (Interactive Questioning, Ques): Actively prompts participants to reflect on whether the meeting is deviating from its goals by asking questions at critical moments.

• What are the innovative aspects of this solution?

  1. Applying two distinct design approaches—passive and active reflection—to meeting management.
  2. Integrating real meeting data and user contexts to enhance ecological validity.
  3. Focusing on the content (What), timing (When), and roles (Who) of reflection, and proposing ways to optimize AI interventions through personalization and timing adaptation.

• What are the implementation steps and key technologies used?

  1. Collecting and processing real meeting data: Including meeting recordings and transcribed notes.
  2. Embedding reflection technology probes:
     • The passive system uses a streamlined visual interface to display topics and content related to meeting goals in real time.
     • The active system inserts guiding questions at critical discussion moments to prompt immediate reflection.
  3. User experience and interview feedback: Using the Video-Stimulated Recall method, participants imagined their actual reactions to these tools.
  4. Data analysis: Conducting thematic analysis of user feedback and summarizing current practices and future design recommendations based on meeting records.

Research Outcomes

• What specific outcomes were achieved?

  1. Users valued the role of meeting reflection in clarifying and adjusting goals.
  2. The passive tool (Viz) effectively promoted non-intrusive goal awareness, while the active tool (Ques) was better at triggering immediate reflection and action.
  3. Three major design dimensions and practical implementation suggestions were proposed:
     • What (Reflection Content): Should combine descriptive, contextual, analytical, and actionable information.
     • When (Reflection Timing): Includes objective timing (e.g., at the start or end of a meeting) and subjective timing (e.g., when discussions deviate from goals).
     • Who (Reflection Target): Provide customized information based on different user roles (active participants, passive participants, and meeting organizers).

• What advantages does it have compared to existing solutions?

  1. Higher ecological validity: Using real meeting data makes the research findings more applicable to real-world work scenarios.
  2. In-depth exploration of the strengths and weaknesses of passive and active reflection, as well as their adaptation to different user roles and situational needs.
  3. Offers strategies for balancing inclusiveness and efficiency in team collaboration through reflective practices.

• What were the experimental or evaluation results?

  1. The passive system is better suited for non-disruptive task awareness, though its cues are occasionally overlooked.
  2. The active system can trigger immediate reflection but may cause interruptions and cognitive load, requiring dynamic adjustment of its intervention intensity based on specific scenarios.
  3. Users expressed significant needs for information richness, low error rates, and controllability, while emphasizing the importance of timeliness and adaptability to different meeting types.

• Limitations and Future Directions

  1. Limited sample diversity (participants were all from a single tech company); future studies could include users from different industries and cultural backgrounds.
  2. The current study was conducted in a pseudo-real-time (pre-recorded playback) environment; future research should deploy technology probes in real-time meetings to validate their practicality.
  3. Further optimization of AI feedback content quality and real-time mechanisms is needed, along with improved modeling of functional requirements for different user roles.

By combining the distinct characteristics of passive and active reflection and systematically considering design dimensions, the authors have taken a significant step toward supporting efficient goal-oriented meetings. This research provides profound design insights for the development of meeting reflection technologies and offers broad implications for the field of intelligent collaboration systems.