Solution Briefs

5 www.radisys.com RADISYS SOLUTION BRIEF | Integrated Speech Recognition Solution Powers Next-Gen Service Providers The Radisys Engage Media Server integrated "in-call" keyword detection and media analytics solution creates new opportunities for service providers by enabling cost-effective and new innovative applications including IVR, enhanced customer service, personal digital assistants, real-time advertising, information retrieval, network assistants and more. Use Cases for "In-Call" Speech Recognition Use Case 1: Reducing Cost of Voice-Enabled Call Center Deployment with Embedded Keyword Detection As CSPs and other operators work on reducing customer support costs, speech-enabled IVR is often used to handle calls, provide 24x7 services, improve customer satisfaction and increase productivity. However, using ASR solutions with full NLP capabilities for all speech recognition needs is not cost effective. While advanced full natural language processing and analytics capabilities are required to optimize the user experience and offer enhanced services completely, a large percentage of speech recognition entails processing only a limited set of keyword vocabulary. CSPs and SIs can significantly reduce the cost of implementing speech recognition in call center solutions with keyword detection system embedded in real- time communication media server. Just like the workflow introduced by smart speakers with an embedded small footprint of speech processor, media-engine integrated keyword detection can be used to do an initial screening of keywords and invoke more advanced NLP, Text-to-Speech, and analytics solutions on an as- needed basis based on business rules. This reduces the hardware required to handle large volumes of calls, and SW license costs. Also, with low-cost "in-call" embedded keyword detection function in a media server, new voice-enabled call-center applications can be created and extended to a broader market. Use Case 2: Conversation Analytics for Personalized Marketing and Advertising One of the areas where "in-call" keyword detection can make an impact for communication providers is to track certain words and phrases in conversations (real-time and recorded) that would help retailers and marketers to build precise target advertising. The speech analytics element in the solution can further build consumable marketing data that can create a revenue stream for the operators as well. Use Case 3: "In-Call" Personal Digital Assistant Services for Peer-to-Peer Calls and Multi-Party Business Conferencing Applications Most of the speech recognition service innovation is currently happening in consumer applications that have a dependency on a smart speaker that connects to a cloud via the Internet. Though use of smart speaker as personal digital assistant to listen to music, get general answers, and control smart home devices will continue to grow and become part of our everyday lives, accessibility to a device independent personal digital assistant, within a context of a phone conversation, remains an open gap for both individual and business users for two-party and multi-party conferencing calls. As per 2018 NPR (National Public Media) report, 43 million Americans own a smart speaker. In contrast, over 235 million Americans own a smartphone and spend on overage 646 minutes per month chatting on the phone . CSPs have a tremendous opportunity to disrupt the market by incorporating in call voice- enabled personal digital assistance services within their consumer voice and business communication and collaboration services such as UC.

• Cover