AI-Powered Conversational Interfaces
summary
Nuance demos an enterprise pipeline that goes beyond buzzwords: an intent discovery process that clusters chat logs and auto-labels intents, cutting manual labeling from hundreds of hours to days and bootstrapping models at ~65–70% accuracy with a human-in-the-loop “hidden agent” for unknowns. The stack includes IVR+chat convergence, a targeting engine for when to pop a bot, a product that ingests websites for question answering, and experimental voice biometric and behavioral biometric authentication—blocked on Alexa/Google by audio-for-text APIs and constrained by strict customer data/privacy rules.
AI-written summary · about
chapters (18)
- 0:00Intro
- 2:10Pauls background
- 4:32What is Nuance
- 6:25What is IVR
- 7:59Title of Pauls talk
- 10:32Voice User Interface Design
- 12:46Biometrics
- 13:43Unsupervised Learning
- 14:55Unsupervised Methods
- 16:51Business Model
- 19:33Customer Model
- 24:32Standards
- 26:08AI and Machine Learning
- 27:55Ontology
- 29:22FAQ
- 31:00Transfer Learning
- 32:01Identifying Intent
- 35:14Conclusion
related talks
23mContact Center Voice AI: Low-Latency Intelligence Extraction from Messy Audio Streams
47mBuilding sophisticated conversational AI experiences
52mAI Powered Contact Center Analytics
37mRole Of AI In Consumer Interaction
34mCognizant Intelligent Interactions: Elevating your customer experience
19mThe Advantages of Building Your AI Stack In house
30mConversational AI: Building clever chatbots
01h 04mBuilding Expertise Into Generated Conversations