『How to Get to “Where’s My Package” | Ep. 26』のカバーアート

How to Get to “Where’s My Package” | Ep. 26

How to Get to “Where’s My Package” | Ep. 26

無料で聴く

ポッドキャストの詳細を見る
One in two calls never reaches a human agent. This statistic highlights a shift from using AI strictly for routing to using it for true resolution. Jon Zoltie joins Oren Batchelor to discuss the mechanics of resolving high-volume requests—like "where is my package"—without human intervention. The conversation explains how Large Language Models (LLMs) and APIs work together to perform data queries, check order statuses, and even use external data, such as weather reports, to explain delays.ㅤThey also address the dangers of the "Super Bot"—an attempt to automate sales, billing, and support in a single unmanageable system. Oren explains why specialized agents are more effective and details how to establish guardrails to prevent AI from hallucinating or accessing sensitive data. The discussion concludes with practical steps for leaders to help identify call topics suitable for automation.ㅤGuest BioOren Batchelor is a Contact Center Specialist at AudioCodes, working within the UK team to bridge the gap between technical capability and business requirements. He focuses on migrating and operationalizing contact center solutions within the Microsoft Teams ecosystem, specifically handling the Voca CIC line of business.ㅤWhat We CoverThe technical process behind AI resolution includes intent detection and API lookups for real-time tracking.Using external data reasoning (e.g., snowstorms in Germany) to explain delivery delays to customers.The "Human in the Loop" concept: allowing AI to process small refunds while routing high-value approvals to agents.Why high-empathy scenarios, such as grief or bereavement, must always bypass automation.The seamlessness of modern insurance policy updates compared to legacy voice interactions.The efficiency gains of removing mundane tasks like password resets from an agent's workload.The "Super Bot" mistake: Why trying to build one bot for every department often fails.Practical first steps: analyzing the last six months of call data to find candidates for automation.ㅤResources MentionedAudioCodesVoca CICGoogle AI Report 2026ㅤGet Started with Voca CIC👉 Start a free trial of the Voca Conversational Interaction Center for 30-days: https://voca.audiocodes.io/signup👉 Learn more about AudioCodes Voca CIC: https://www.audiocodes.com/solutions-products/voiceai/voca-conversational-interaction-center
adbl_web_anon_alc_button_suppression_c
まだレビューはありません