Who should be my teammates: using a conversational agent to understand individuals and help teaming
summary
An agent named INDIGO on Juji conducts text interviews twice in a semester with 184 students, 40 teams, collecting free-text answers and automatically inferring personality facets; students average 60 minutes pre-project interview and 620 words, 20 minutes and 289 words post. Regression models show higher variance in emotional traits -> better team performance, and lower sensitivity predicts less conflict. Students rate the agent likable and trustworthy but not very enjoyable; the system can recommend team composition from inferred traits.
AI-written summary · about
chapters (11)
- 0:00Introduction to INDIGO
- 1:54Challenges in team formation
- 2:36INDIGO system overview
- 4:23Study setup and deployment
- 6:09Student interaction with INDIGO
- 7:39Effectiveness of data collection
- 9:06Insights on team personality
- 10:50Future design implications
- 12:18Study limitations
- 13:01Conclusion and thanks
- 13:22Q&A session
related talks
29mThe AI Teammate Is Here — Now What?
49mAI agents won’t save your productivity– human–agent collaboration will
59mFrom Pair to Peer Programmer
15mVibe Coding at Scale: Customizing AI Assistants for Enterprise Environments
26mAI Agents as Your Co-Developer
19mYour Biggest User Isn't Human: Designing DevTools for AI Agents
26mFrom autocomplete to agents: AI coding assistance state of play
35mThe Podcast Pipeline Nobody Sees: Behind the Scenes with AI Agents and AI gen