The Contact Center Gurus From AI Chaos to Customer Confidence: Building Virtual Agents That Actually Work
10 snips
Dec 18, 2025 Adam Fenwick, SVP of Global Alliances at Boost.ai, explores the intricacies of conversational AI in contact centers. He highlights common pitfalls like rushing deployments and inadequate conversation design that lead to failures. Adam advocates for a strategic, multi-channel AI approach and discusses the importance of balancing generative models with accuracy. He shares success stories from telecom and healthcare, showcasing significant ROI and improved customer satisfaction through well-designed virtual agents. His key takeaways emphasize thorough planning and vendor diligence for optimal outcomes.
AI Snips
Chapters
Transcript
Episode notes
Treat Conversational AI As Strategic
- Treat conversational AI as a strategic, company-wide initiative rather than a series of small pilots.
- Narrow POCs chasing 'low-hanging fruit' often block scaling and break the business case over time.
Avoid Narrow, Misleading Pilots
- Avoid starting with a handful of narrow, easy use cases that don't represent all customer topics or channels.
- Start with a broad vision (all channels, all languages, all topics) and pilot from that perspective.
Voice Needs Different Design Tradeoffs
- Voice is higher-risk and higher-expectation than chat, so design and architecture choices matter more for voice.
- Different voice AI architectures trade off speed, naturalness, accuracy, and security in distinct ways.