For CSPs, there are four clusters where artificial intelligence (AI) can and will play a role in enhancing operations, supporting innovation and driving growth, and where the industry ecosystem (CSPs, vendors, regulation and standardization bodies, analysts, etc.) is already working at identifying relevant use cases and designing AI-powered solutions.
1. Customer engagement is the most advanced area, although far from mature. CSPs have in this area, a lot in common with other industries, specifically those which offer complex products and services to wide customer bases such as insurance and banking. Cross-fertilization in innovation and implementation is therefore of the essence.
There are specifically two components in customer engagement where AI can make a real difference: customer care and digital marketing.
In customer care, virtual assistants (chatbots, voice assistants, smart speakers, conversational platforms) are already broadly introduced and some global CSPs have even developed their own Alexa-like platform, e.g., in Europe, Orange with Djingo (in cooperation with DTAG), Vodafone UK with TOBi (“the first live chatbot in UK telecoms”), Telefonica with Aura (“Telefonica’s Artificial Intelligence”) and many others.
The footprint of virtual assistants will quickly expand in telecoms as well as in other industries. Many analysts predict that within a few years there will be more virtual assistants installed across the globe than human beings and a significant part of the customer engagement processes will be through virtual assistants, compared to around just one per cent today.
Nevertheless, human interaction will remain a key differentiator in customer care, as highlighted recently by DTAG CEO Timothy Hoettges, who noted that “Having a haptic feeling, talking to someone in flesh and blood will help us to stand apart”.
One simple way CSPs can leverage the best of both worlds is to implement AI tools to support call center agents and that is feasible now.
In digital marketing, AI tools can enhance customer segmentation, real-time personalization, next best offer/action and customer journey insights. Beyond global vendors’ value propositions such as Salesforce’s Einstein, many other players are also developing generic or niche solutions.
Loyalty management is another customer engagement track that AI platforms can take to the next level. For example, Telefonica is pairing its loyalty application Priority with its Artificial Intelligence Aura to provide loyalty advantages on personalized opportunities identified by Aura.
2. Network management and optimization is where CSPs will get the biggest benefits from AI, at least in the longer term. Because of the complexity of network environments and because the emergence of AI is concurrent with other drastic network transformations such as virtualization and 5G, it will take some time for our industry to be able to design and roll-out telco-grade solutions.
While probably the less mature area for AI implementation, the impact of AI on network operation enhancements and the resulting cost reductions are potentially huge and, beyond these clear benefits, AI will be needed to manage the complexities of next generation networks.
The industry is already working at implementing and leveraging cognitive technologies in these three areas:
- Traffic management with solutions to categorize services automatically based on traffic characteristics analysis, and to subsequently define the proper level of network resources for each service.
- Automation of software-defined networking (SDN) and network functions virtualization (NFV) operations for complex events correlation, root cause analysis, preventive maintenance, preventive assurance and incident resolution; in virtualized as well as in hybrid networks.
- Self-organizing network (SON) technologies to ensure self-configuration and self-optimization of mobile radio access networks, and in the longer term, to move to self-maintenance and self-healing, where AI can drive responsiveness and efficiency.
While vendors are heavily investing in AI-driven network management, operators and industry organizations are also leading some major initiatives. Telefónica is currently working with Juniper Networks to identify uses cases, tools and processes to leverage AI/machine learning to manage and optimize their network.
In 2017, ETSI created the industry specification group ‘Experiential Network Intelligence’ (ISG ENI) to define a cognitive network management architecture which uses AI techniques.
And, of course, several groups in TM Forum are actively working at shaping the future AI-enabled digital operator.
3. Service and business assurance is where AI technologies can significantly enhance the performance of current solutions and open the door to disruptive approaches.
In service assurance, where the human experience of recurring faults and incidents correlation is a key element for fast and efficient resolution of trouble tickets, machine learning techniques can play a major role in the automation of the assurance processes and the ‘zero-touch assurance’ becomes a realistic objective thanks to AI. That also includes prediction of networks faults and failure prevention, where AI solutions can propose rules and criteria to engage preventive maintenance. The broader development of closed-loop automation (i.e. without human intervention) enabled by AI leads to the transformation of network operations centers into services operation centers, a concept that Telefonica is testing in different countries.
AI is naturally also fully applicable to fraud detection and prevention and, therefore, a key element of revenue assurance solutions, which will leverage more and more cross-industry algorithms as this threat in CSPs operations is very similar with what other industries face. AI is likely going to be embedded in most of the next generation revenue assurance solutions.
Cybersecurity, a major risk for society at large, and highly sensitive for CSPs in several ways. This includes protecting their own IT and networks against cyber threats and containing potential cyberattacks, but also the opportunity to offer cybersecurity as a service to their corporate clients, and, maybe in the future, to their consumer base. The potential of AI in cybersecurity is unquestionable and supported by a wide spectrum of players, from institutions like MIT with their AI-based cybersecurity platform (AI2) to a broader community of younger companies focused on this topic.
4. Smartphones and smart devices will likely incorporate AI technologies heavily. As image processing and natural language processing are two major AI-enabled technologies, there is a clear technical and functional rationale to incorporate AI features behind smartphones cameras, micros and speakers. This will open the door to even smarter phones and devices, able to understand their owners and feel their emotions; that will be a true area for differentiation.
Beyond this technical and functional logic, there is a business rationale for the manufacturers to speed up the incorporation of AI into their devices. For the first time, smartphone sales have dropped in 2017 and, in spite of an expected rebound this year, the industry remains under pressure and new disruptive features are badly needed to revitalize the market and maintain growth.
In this context, while manufacturers will be the main drivers in the likely massive incorporation of AI features into smart devices, CSPs have got a strong role to play as well. Not only because they are pushed to do so by the manufacturers’ community or because they re-sell these devices to their customers, but also because the full power of AI in smartphones will require specific interactions with the network and the services delivered by CSPs.
CSPs have a key role to ensure adoption of these technologies by their customers, including helping them to rationalize the questions, threats and myths around the growing intelligence of their smartphones.
Customers will feel comfortable to use these AI features only if there is a joint mobilization of manufacturers and CSPs to educate and reassure them. This intimate collaboration will be key to ensure adoptions of AI in personal devices and wow their common customers.
Beyond the above thoughts on the potential of AI and the description of some relevant use cases, the journey to AI implementation includes the need to address several transversal questions and challenges. These include: governance, talent gap, new services and new business model opportunities, cross-fertilization with other industries, customers trust, societal concerns, etc. These questions will be naturally part of the many debates at digital transformation world, where I hope we will meet many of you.