Online chat is an emerging channel for discussing community problems. It is common practice for communities to assign dedicated moderators to maintain a structured discussion and enhance the problem-solving experience. However, due to the synchronous nature of online chat, moderators face a high managerial overhead in tasks like discussion stage management, opinion summarization, and consensus-building support. To assist moderators with facilitating a structured discussion for community problem-solving, we introduce SolutionChat, a system that (1) visualizes discussion stages and featured opinions and (2) recommends contextually appropriate moderator messages. Results from a controlled lab study (n=55, 12 groups) suggest that participants' perceived discussion trackability was significantly higher with SolutionChat than without. Also, moderators provided better summarization with less effort and better managerial support using system-generated messages with SolutionChat than without. With SolutionChat, we envision untrained moderators to effectively facilitate chat-based discussions of important community matters.


  • Sung-Chul Lee
  • Jaeyoon Song
  • Picture of Eun-Young Ko Eun-Young Ko
  • Picture of Seongho Park Seongho Park
  • Picture of Jihee Kim Jihee Kim
  • Picture of Juho Kim Juho Kim

Main screen

(A) agenda panel for showing the discussion structure, (B) current discussion stage and featured opinions, (C) stage divider, (D) button for add a message as a featured opinion, (E) inline message recommendations for moderators to add short reactions to discussants' opinions, and (F) block message recommendations with generic facilitation messages for moderators to use.

Message Recommendation Design

SolutionChat recommends moderator messages based on the information of Agenda Panel and NLU classification. If participants use Agenda Panel, Agenda Panel can capture the featured opinions, current stage and vote status. Based on those informations SolutionChat recommends managerial messages for stage management and other pedagogical messages for argument developing activity. Also, SolutionChat classifies each discussants and moderator messages with the NLU module. For the discussants, NLU is trained to detect discussants' opinion suggestion message (e.g. I think ~). Once NLU detected a suggestion message, SolutionChat recommends Inline MR for a moderator's social and pedagogical supports. A moderator can take a recommended message by clicking the recommended message or type it manually. For a later case, NLU can match manually typed recommended message and dismiss the corresponding recommended message automatically.



CHI 2020 Paper (Camera-Ready)

Camera-Ready @ KIXLAB.org


CHI 2020 Slides


Video Presentation


CHI 2018 LBW paper

Camera-Ready @ KIXLAB.org

@inproceedings{lee2018micro, title={Micro-NGO: Tackling Wicked Social Problems with Problem Solving and Action Planning Support in Chat}, author={Lee, Sung-Chul and Kim, Jihee and Kim, Juho}, booktitle={Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems}, pages={1--6}, year={2018} }


CSCW Workshop paper



This work was supported by the Office of Naval Research (ONR: N00014-18-1-2834), and by Institute of Information & Communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No.2016-0-00564, Development of Intelligent Interaction Technology Based on Context Awareness and Human Intention Understanding).