White Paper AIOps – Automated Dissatisfied CX Analysis May 2019 Mark Geere, Jing Hao, Chiu-Chung Yue, Derek Chen HUAWEI SmartCare® HUAWEI TECHNOLOGIES CO., LTD. Copyright © Huawei Technologies Co., Ltd. 2019. All rights reserved 1. Executive Summary Artificial Intelligence and Machine Learning has spread to all industries and all parts of those industries. Whether we are talking about Enterprise adaption within the Automated Cars, the Health industries etc. or within the consumer world where it is now embedded within everyone’s phones to help them take a better picture or more efficiently use their battery power. The point is that AI is now ubiquitous throughout our lives at work, home and within the products and services we consume. So it is no surprise that the Telecoms industry is also looking at how they can use AI to (1) improve internal efficiencies across their organizations, as well as (2) drive better customer engagement to enhance customer experience on the services they deliver (Figure 1). The “Automated Dissatisfied CX Analysis” use case is a part of the AIOps catalyst which links Hong Kong Telecom (HKT) with Huawei’s SmartCare® Customer experience solution to deliver on these 2 key areas. The focus of the use case is to investigate how customer complaints at the point of reporting can be intelligently analyzed in order to deliver the following benefits: Support network optimization: By aggregating affected users and problem dimensions, this use case supports network problem optimization based on user experience assurance and improves O&M efficiency. Support proactive care: Provides proactive care for identified affected users to improve customer satisfaction. Support complaint handling: Is the preprocessing of features from complaining subscribers so that when a subscriber complains, the solution can quickly obtain the abnormal characteristics of the subscriber and locate the fault. 2. Telecoms Industry AI Drivers TMForum’s recent insight report “AI and its Pivotal role in Transforming operations” found that, out of the 69 operators they contacted, the top 2 primaryHUAWEI drivers for the use of AI was SmartCare® to (1) Drive better customer experience and (2) Reduce OpEx in the organization by driving automation and closed looped systems (Figure 1). Copyright © Huawei Technologies Co., Ltd. 2019. All rights reserved Figure 1 - AI and its Pivotal role in Transforming Operations The report also indicated where the areas of spend that the CSPs are looking at when they focus on AI as part of their Transformational (Figure 2). The main focus was looking at trying to predict and fix issues within the network that support customer care prior to those complaints escalating. Figure 2 - AI and its Pivotal role in Transforming Operations HUAWEI SmartCare® This is why HKT and Huawei collaborate within the AIOps Catalyst to focus on a use case that makes use of customer complaints to define and predict where issues within the network will cause severe impact and proactively fix them before they deteriorate to major issues. Copyright © Huawei Technologies Co., Ltd. 2019. All rights reserved 3. Business Requirements With over 60 million annual tourists from both local residence and roamers visiting Hong Kong, one of the main reasons HKT decided to join the catalyst was to deepen its customer engagement through enhanced customer experience for smarter business operations in order to offer a unique, integrated quad-play experience. The company wanted to be able to detect abnormal customer experience early and fix these issues before they become more serious complaints. Within the catalyst, HKT intended to explore how AI algorithms can analyze and collect performance data to drive predictive and corrective actions which would minimize non-effective cell site visits and identify hidden problems that were causing customer complaints. To this end this AIOps Catalyst use case was created to look for a solution that could proactively use customer data to: 1. Aid the frontline call agents to understand whether its customers were potential dissatisfied customers prior to their calling. 2. Reduce the number of complaint calls by being able to focus on areas of high impactful complaints. 3. Enable network operations and optimization departments to quickly and efficiently target fixes to trouble ticket impacting areas. 3.1. Why use AI for network related complaints Notwithstanding the above points, HKT was keen to ensure that they comply with all customer data privacy restrictions. Therefore HKT was looking for method that would use the insights from the customer complaints and the corresponding data sets without falling foul of any consumer data restrictions. 4. Huawei and HKT AIOps Solution HKT decided to partner with Hauwei’s SmartCare® solution to investigate and create a proof of concept that would deliver a catalyst to investigate how customer complaints, at the point of reporting, can be intelligently analyzed and then to deliver integrated value to both customer care and network operations. 4.1.HUAWEI Solution Overview SmartCare® Huawei used its extensive knowledge in AI and CEM to create a proof of concept along with HKT that used AI and CEM data sets combined with CRM data to deliver process transformation that meet HKT’s business requirements. Copyright © Huawei Technologies Co., Ltd. 2019. All rights reserved Figure 3 The solution is categorized into 4 steps (Figure 3): Step 1 – Dissatisfied User Classification: Offline training process takes the user complaint analysis to build a classification model, then train the model to understand and then discover potential dissatisfied customers’ experience. Step 2 – Anomaly Clustering: In production process (In Production stage) the AI model automatically discover the dissatisfied customers, identifies the anomalous patterns affecting their user experience and the corresponding problems Step 3 – Actionable Insights: By changing the processes within the organization HKT are able to integrated different parts of the organizations and provide insights to CRM team as well as demarcation information for operations and optimization teams. Step 4 – Impact Monitoring: SOC then provides a continuous monitoring functionality to ensure that the network change has had a beneficial impact on the customer experience indicators and therefore the customers’ satisfaction. Key AI Algorithms: • Deep learning: to discover the distinguishing characteristic of the experience indicators of the complainants. • Sequential Pattern Mining: to reveal the abnormal sequential event patterns of the complainants. • Classification modelling: to identify the abnormal users based on HUAWEImultiple data patterns. SmartCare® 5. Business Benefits As this catalyst project is still ongoing, actual benefits have yet to be fully quantified but initial business value modelling indicates that the following business benefits should be achievable: Copyright © Huawei Technologies Co., Ltd. 2019. All rights reserved Corporate Brand Improvement: Improve HKT Brand reputation for delivering the best quality service by being proactive in determining issues that will lead to customer complaint. Targeted to improve NPS by 5~10 points OPEX and CAPEX Efficiency improvement Reduce the number of calls by 10% Improve AHT by 10~15% Improve Mobile services EBITDA margin by 3% Process Improvement Value Improve HKT process to proactively identify & demarcate network issues and optimize inter-silo cooperation from CRM to Operations and Optimization. Predicting to Gain 5% improvement 6. Next Steps The catalyst project would aim to: 1. complete a fully integrated run of the initial solution to show actual business benefits 2. improve the offline training process to determine proactive knowledge base to be more proactive through the “Network Problem Mode Training Process” 3. provide a business model to show tangible ROI results Key Project Members Name Email Champion Chiu-Chung Yue [email protected] Derek Chen [email protected] Rudy Wong [email protected] Participant Mark Geere [email protected] Jing Hao [email protected] Huiru He [email protected] HUAWEI SmartCare® Copyright © Huawei Technologies Co., Ltd. 2019. All rights reserved About HKT HKT (SEHK: 6823) is Hong Kong's premier telecommunications service provider and leading operator in fixed-line, broadband and mobile communication services. It meets the needs of the Hong Kong public and local and international businesses with a wide range of services including local telephony, local data and broadband, international telecommunications, mobile, and other telecommunications businesses such as customer premises equipment sales, outsourcing, consulting, and contact centers. HKT offers a unique quadruple-play experience in Hong Kong delivering media content on its fixed-line, broadband Internet access and mobile platforms jointly with its parent company, PCCW Limited. HKT also provides a range of innovative and smart living services beyond connectivity to make the daily lives of customers more convenient, whether they are at home, in the workplace, or on the go. Employing over 17,400 staff, HKT is headquartered in Hong Kong and maintains a presence in mainland China as well as other parts of the world. For more information, please visit www.hkt.com. About Huawei Huawei is a leading global provider of information and communications technology (ICT) infrastructure and smart devices. With integrated solutions across four key domains – telecom networks, IT, smart devices,