Dynamic storytelling and NPC interactions using generative AI & Azure Cosmos DB in game development
summary
Shows a working prototype that stores Azure Cosmos DB V4 documents with OpenAI embeddings and vector indexes (HNSW/IVF/DIS) to power RAG-style NPC dialogue. Demo pushes name/message/dialogue vectors, queries with a relationship filter so the same prompt returns hostile/friendly/neutral NPC responses per player profile. Also cites a KPMG implementation where search accuracy rose from ~50% to 91%. Implementation details include vector dimension handling, using Cosmos as the vector store, and a live demo that successfully changes NPC replies based on stored relationships.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 1:25Azure Cosmos DB
- 4:08Azure Cosmos TB
- 6:58NPC Interaction
- 9:10Create Story
- 17:52Summary
related talks
46mMind your NPCs and Q’s: what’s next for Game AI?
50mBuilding AI ready applications
23mForget everything you think you know about NPCs and world-building
50mA fun and absurd introduction to Vector Databases
51mGame of Bots
48mA fun and absurd introduction to Vector Databases
46mA Fun & Absurd Introduction to Vector Databases
12mThe mechanics of integrating AI and vector search into Azure Database for PostgreSQL