Cognitive Contact Center Catalyst Project May 14 17

Cognitive Contact Center Catalyst Project May 14 -17, 2017 – Nice, France © 2018 TM Forum | 1

Market Drivers The 2020 CSP: assuring the engagement with customer Telecom customers expect to duplicate the use of digital channels but at the same time, they consider the Contact Center as the preferred channel for customer care interaction in the future Current use Future Preference Shop 63% Contact Center 46% Information Search Service purchase 43% 20% 21% 10% 11% 25% 22% 19% 15% 56% Service purchase 43% Of the customers expect to interact with the Contact Center during post-sales Customer Care Digital Channels 50% 59% Information Search 15% 33% Service purchase Current Use 19% 35% Customer Care Future preference SOURCE: Connected Telco LATAM, market Research done by everis based on 3600 interviews to telecommunication users, 2017 © 2018 TM Forum | 2

Market Drivers The 2020 CSP: ensuring the engagement with customer DURING post-sale, customers seek care in every moment and agility; this puts a demand on CSPs to be proactive and to anticipate problems What aspect do you value the most of the customer care? Main causes of customer interaction: 23% To be able to communicate with MY TELCO in every moment Problem resolution in few steps Honesty in regards of the problem reason and resolution time Temporal solutions while my problem is solved Possibility to give feedback about the service 39% 23% 42% 22% 21% 23% 26% 40% Bill related request Technical problems 34% 37% SOURCE: Connected Telco LATAM, market Research done by everis based on 3600 interviews to telecommunication users, 2017 39% © 2018 TM Forum | 3

Cognitive Call Center: vision We envision a Cognitive Contact Center that truly engage native and non-native digital customers enabling a complete omnichannel digital customer care ALWAYS AVAILABLE COGNITIVE (no menus) NO QEUES PREDICTIVE SENTIMENT ANALYSIS PERSONALIZED SMART ASSISTANT END TO END ANTICIPATE NEEDS Emotional connection Predictive actions ANALYTICS EASY TO USE Greater recurrence Personalized experience COSTS REDUCTION USEFUL Information and transactions Functional satisfaction © 2018 TM Forum | 4

Cognitive Call Center: approach Our approach for the Cognitive Contact Center is based on the continuous and progressive implementation of use cases to be moved to an omnichannel Virtual Agent AS IS TO BE – Cognitive Contact Center Customer Contact Center Agent Management Systems Contact Center Agent Management Tools & Systems Unified Service Desk Omnichannel Virtual Agent Intelligence Conversational Context © 2018 TM Forum | 5

Cognitive Contact Center: High Level Architecture The architecture components of this initial Catalyst phase are totally connected and it could be possible to deploy them in the short term. In fact, it has been possible to enable them leveraging existing implementation across Telefonica Group operations Engagement Layer Traditional Channels Contact Shop Center Digital Channels Mobile Web Mobile App … Voice Recognition & Text To Speech Voice Recognition Cognitive Layer Everis Virtual Agent (e. VA) (Broker & Cockpit) Text To Speech everis Virtual Agent Cognitive Services Sentiment Analysis Integration Layer API Management & Portal Core Layer Service Virtualization B/OSS Service Management Platform Unified Service Desk CRM © 2018 TM Forum | 6

Cognitive Call Center: TM Forum OPEN APIs 8 TM Forum OPEN API supported TMF Open API’s 1 Party Management REST API GET /PARTYMANAGEMENT/INDIVIDUAL 2 3 Traditional Channels Contact Shop Center Product Catalog Management REST API Voice Recognition Product Inventory REST API Everis Virtual Agent (e. VA) (Broker & Cockpit) Billing Management REST API Service Problem Mgmt. REST API GET /service. Problem/ 7 Usage Consumption REST API GET /usage. Report 8 Product Order Mgmt. REST API POST /product. Order Text To Speech everis Virtual Agent Cognitive Services Sentiment Analysis Integration Layer GET /applied. Customer. Billing. Charge 6 … Voice Recognition & Text To Speech GET /resource GET /product 5 Mobile App Resource Inventory REST API GET /product. Offering GET /product. Specification 4 Digital Channels Mobile Web API Management & Portal 1 2 3 4 Service Virtualization 5 7 6 B/OSS Service Management Platform 8 Unified Service Desk CRM © 2018 TM Forum | 7

© 2018 TM Forum | 8

Verbio Speech Recognition & TTS Verbio helps your bot to TALK Speech Recognition & Transcription Text to Speech Voices We provide all the capabilities for 3 rd party COGNITIVE PLATFORMS be able to communicate using the Voice Channels © 2018 TM Forum | 9

Verbio Speech Recognition & TTS Speech Recognition & Transcription: We understand what the HUMANs say The Verbio’s ASR and CSR engines (Continuous Speech Recognition) have been developed with the most advanced technology and can be used in most of the environments where humans communicate with voice (telephony, app, multimedia, etc. ) ASR / CSR (The EARS) >90% accuracy depending on acoustic conditions Multiple language coverage Deep Neural Networks (DNN) + HMM Noise robust algorithms and trainings Customized to the customer environment. Voices sensors of Voice, Gender and Stress. Text To Speech Voices: We say what your system thinks Our TTS engines enable any system the possibility to SAY anything with a “HUMAN TOUCH” and with the expressivity needed in each moment. TTS (The Voice) Natural voices with intonation and emotions Multiple language coverage Deep Neural Networks (DNN) + HMM Multilingual voices; use different languages in the same conversation with the same voice. Prosody management and dynamic intonation. © 2018 TM Forum | 10

Verbio Speech Recognition & TTS BASIC INTEGRATION ARCHITECTURE: Our engines integrate in multiple ways to the customers environment. IVR (Genesys, Cisco, Avaya, Asterisk, etc. ) MRCP* API / WS Chatbots Verbio ASR Verbio CSR Verbio TTS *We also support other kind of connectors like Webservices, APIs, Websockets, SDK, etc. © 2018 TM Forum | 11

everis Virtual Agent Architecture e. VA architecture is based on decoupled modules and allows customers to integrate easily with their IT stack. • • Logical architecture design • e. VA will solve informational and transactional attendance through multiple channels: Mobile (Native/Hybrid), Web responsive, Social Media, VR/AR, Call center IVR, Voice to text channels as Alexa and others. The solution will provide a multi-channel experience. Any answer to a particular intent can be managed and addressed to a different Channel. e. VA broker is the commissioned layer to manage all messages from chatbot. e. VA API manager is the commissioned layer to standardize, centralize and secure all integrations between e. VA and back-ends systems. The Cognitive Engine is the module of the solution that will be able to receive the information, interpret customer requests, process these requests and share the conclusions with a conversational flow module in order to include them into the system choices. © 2018 TM Forum | 12

everis Virtual Agent Cockpit • e. VA provides a control panel, called cockpit, to manage all knowledge bases, and to view reports through graphical dashboards. • All reports can be exported to excel formats. • Currently e. VA is integrated with Qlik. Sense and Tableau, but it is possible to develop a custom integration with other BI platforms. © 2018 TM Forum | 13

Sentiment Analysis This module gets sentiment information from the user to enrich the conversational context. Sentiment Process • The module will be integrated with several systems that will analyze the text by their own. • A service will receive a string from the broker and return all analyzed responses from the APIs in the same JSON object. © 2018 TM Forum | 14

CA Technology API Management Mobile App Services Internet of Things Enterprise Firewall API Gateway Mobile API Gateway ID/Authentication Live API Creator Mobile SQL Data Saa. S/Cloud Solutions AWS, Google, SFDC … Existing APIs Partner Ecosystems No. SQL Data API Developer Portal Developers AWS/Cloud © 2018 TM Forum | 15

CA Technology CA Service Virtualization • A method of capturing and simulating the behavior of unavailable or incomplete systems Or simulating TM Forum Open APIs • Takes a modern approach to the practice of “mocking” or “stubbing”. • Uses sufficiently “real” virtual services without depending on complexity and constraints of real hardware or software. © 2018 TM Forum | 16

Nokia Service Management Platform Service provider’s pain points in Customer Care Processes for troubleshooting telecom service issues are complex Multiple customer touchpoints resulting in siloed implementations Customer frustrated as each channel has a different experience Customer complains of having to repeat themselves Low adoption of self-service and goal of fixing issues proactively High costs related to repeat calls Operators looking for open platform to do their own development (Dev. Ops) Optimizing troubleshooting processes is manual and tedious © 2018 TM Forum | 17

Nokia Service Management Platform Customer Care using Nokia SMP Nokia Service Management Platform (SMP) is the core of HADA Field proven Omni-channel Customer Care platform for Service Providers CRM Powerful guided care for Agents and end users, complementing existing CRM systems Test & Diagnostics framework to automate troubleshooting Purpose-built tools make it easy for IT developers and Business Analysts Flexible Integration framework to gain end-to-end visibility Pre-built solutions for telecom customer care; HSI, IPTV, Vo. IP, Wifi, Mobile Integrated consoles for CSR Agents, Field Technicians, End users and Enterprises © 2018 TM Forum | 18

Nokia Service Management Platform Benefits of using Nokia Autonomous Customer Care Processes for resolving telecom service issues are not trivial. Avoid duplication by reusing process logic and integrations across channels (Call Center, Retail stores and Chatbots) Single platform to deliver specialized technical support on both traditional support channels and new digital conversational touchpoints Context sharing and seamless continuity of interactions across all channels Self- optimized Machine learning not just in natural understanding layer (NLP) but also in problem identification and problem resolution ML models continually updated via support interactions on all channels including Call Center, Self-service portal and conversational platforms Omnichannel Machine learning Reduce Opex Common, consistent design tools for managing business logic increases synergies across functional groups Extensibility Open platform that embraces and empowers developers Architecture agnostic of NLP engine. Choose NLP engine that provides the best understanding for a language REST API enables developers to build new conversational applications © 2018 TM Forum | 19

Microsoft Dynamics CRM + Unified Service Desk for Dynamics 365 Customer Engagement provides a configurable framework for quickly building applications for call centers so that agents can get a unified view of the customer data stored in Dynamics 365. You can aggregate customer information from different areas in Dynamics 365 into an integrated desktop that provides a 360° view of the customer interactions. This gives your customer service agents immediate access to business critical information so they can quickly engage with customers and address queries and issues. © 2018 TM Forum | 20

Microsoft Dynamics CRM + Unified Service Desk Offer a unified service experience across self-service and assisted service channels. Understand customer history, preferences and voice of the customer to personalize every interaction. Leverage insights and resources to create individualized and consistent experiences across channels. © 2018 TM Forum | 21

Microsoft Dynamics CRM + Unified Service Desk Self-service & communities © 2018 TM Forum | 22

Microsoft Dynamics CRM + Unified Service Desk But Dynamics 365 is not only Customer Service Knowledge © 2018 TM Forum | 23