Practical AI Next-gen voice assistants
Apr 6, 2021
Nikola Mrkšić, CEO & Co-Founder of PolyAI, delves into the exciting world of conversational AI. He shares insights on the evolution of voice assistants, highlighting the challenges in achieving human-level conversations. The discussion includes how users modify their language when interacting with technology and the importance of dialog management in enhancing user experience. Nikola also addresses multilingual processing challenges and the need for comprehensive training data, providing a fascinating look into the future of voice technology in customer service.
AI Snips
Chapters
Transcript
Episode notes
Single-turn vs. Multi-turn
- Mainstream voice assistants like Alexa and Siri primarily handle single-turn interactions.
- Multi-turn dialogues, crucial for complex tasks, pose significant evaluation and development challenges.
Human-Computer Interaction Quirks
- People interact differently with voice assistants than humans, often using shorter sentences and more profanity.
- Early Alexa apps focused on simple functionalities like animal sounds, highlighting the evolution of technology.
Voice's Strengths and Customer Service
- Voice excels as an interface for on-the-go scenarios and high-bandwidth communication.
- Despite the rise of digital channels, voice remains dominant in customer service, especially during crises.