StorySage: Conversational Autobiography Writing Powered by a Multi-Agent Framework
summary
A system called StorySage implements a five-agent pipeline — Interviewer, Session Scribe, Planner, Section Writer, Session Coordinator — that stores responses in a memory bank and question bank, drafts follow-up agendas, and integrates new material mid-session by having the planner fire every M memories (e.g., 5–10) to create or update sections. A user study (N=28, 14 vs 14) shows higher conversational flow, narrative completeness, and user satisfaction versus a baseline while responsiveness stays comparable; LLM simulations model longer-term use.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:09Meet StorySage
- 3:16Design goals
- 5:43System architecture
- 8:08Biography writing
- 9:55Future session planning
- 10:15User study results
- 11:39Limitations
- 12:21Q&A
related talks
44mBringing stories to life with AI, data streaming and generative agents
18mHow structured thinking gives your AI superpowers
33mOwn Your Story: Turning Generative AI into Human-Sized Narratives
43mDesign Thinking: Generative AI Style
45mAn Agent Walks Into a Bar... and Meets Another Agent: Multi-Agent Systems in Action
49mAI Agents: Collaboration in Real-Life Applications
32mBuilding Multi-Agent Systems: From Concepts to Collaboration
41mBringing stories to life with AI, data streaming and generative agents