Building Expertise Into Generated Conversations
summary
Large language models reshape conversation design by doing concrete work: using few-shot models to generate training utterances and extract voice-and-tone from client websites, analyzing open‑ended research responses, and auto‑producing sample dialogues so designers write far fewer scripts. At runtime they power flexible entity collection (one AI node replaces spaghetti logic), procedural handling, and RAG-based FAQ answers (example: “third checked bag costs $35–$55”). Tradeoffs are hallucination, latency, privacy/compliance risk, and the “stochastic parrot” limitation, so teams enforce prompt engineering, guard rails, and human review.
AI-written summary · about
chapters (9)
- 0:00Introduction and context
- 3:06Impact of generative AI on conversation design
- 8:09AI technology stack in conversational UX
- 17:43Human conversation complexity
- 25:36Using AI in design practice
- 31:15Generative AI at runtime
- 38:21Risks and guardrails for generative AI
- 42:36Stochastic parrots and language models
- 47:05Q&A session
related talks
17mBuilding Effective Voice Agents
25mConversational AI
55mDesigning for speech
58mA Primer on AI Powered Conversational Experiences and Language Models
26mMastering Large Language Models: Prompt Engineering for Precision & Performance
45mPanel Talk
48mMastering the AI Shift: Adaptive Strategies for Software Evolution
35mDesigning Experiences for Chatbots