Artificial intelligence (AI) is the subject of considerable discussion across industry sectors around the world. Telecoms companies are showing considerable interest in the potential of AI, whether at the front end—customer service—or to improve service delivery via better network performance and reliability.
There are a number of different AI technologies that may be useful in telecoms. These include
- Self-optimising networks (SON), where the network’s goals and limits are set by designers, and the network’s control software works within those boundaries to make the network as efficient as possible;
- Deep neural networks may enable machines perform human like tasks, allowing them to help digitalizing the business and providing better customer engagement;
- Software defined networks (SDN) and Network Function Virtualisation (NFV) will increase the diversity of possible traffic through the network. Both services and bundles will be able to become more sophisticated, with customers even able to interact with the services behind the network.
Improving networks and processes
At the moment, telecoms operators are just starting to consider the idea of AI to improve networks. The European Telecommunications Standards Institute (ETSI) recently announced the establishment of a new group to examine how AI might be used to improve the way in which telecoms networks operate. The group is concentrating on Cognitive Management Architecture, to enable the creation of networks that can adjust services based on user need, environmental conditions and business goals. The system will learn from experience, configuring networks to meet demand, and therefore improving network use and maintenance, and reducing costs.
Members of the group include Verizon, Huawei and China Telecom. Verizon was already considering the use of AI, through its Exponent initiative, and AT&T is also experimenting with the concept. AT&T’s Domain 2.0 initiative draws on the potential of software-defined networks, and aims to transform the network and infrastructure is something that more resembles cloud computing. Another example is ZeroStack’s ZBrain Cloud Management, which analyses private cloud telemetry storage and use, improving capacity planning, upgrades and general management.
Telenor, in Norway, has taken a different approach to AI development. It has contributed funding and anonymised data to help establish a new AI research centre, based in Trondheim and linked to the Norwegian University of Science and Technology (NTNU). Research organisations and commercial partners are expected to come together at the centre to work on new AI applications and how they can be used in practice.
Tackling customer services
These initiatives largely address network functioning and particularly efficiency of processes. AI can, however, also be used to improve customer service. For example, several telecoms companies are offering AI systems that can control home environments, manage scheduling and make music recommendations, such as NUGU from SK Telecoms. This sounds a bit like competing with Amazon’s Alexa or Google’s Assistant, but other telecoms companies have focused AI more on their core services. IPSoft’s Amelia, for example, can replace customer service teams, and interacts with customers, learning from experience what actions to take in particular circumstances. To improve the customer experience.
Last month, April 2017, Vodafone announced the arrival of its new chatbot, TOBi, to help customers online, and shortly to be introduced to the My Vodafone app’s messaging service. TOBi will be able to handle a range of customer service-type questions, including troubleshooting, order tracking, and usage. More to the point, it will speed up responses to simple customer queries, therefore delivering the speed that customers want.
O2 announced recently that it, too, would use AI technology for some customer service interactions. Its new platform, Aura, is designed to reduce customer service costs and will be used over the phone. The CEO of O2’s parent company, Telefónica, suggested that Aura will improve customer loyalty and experience by handing back control of the data generated by smartphones to customers. It can also be used for tasks such as informing a bank that you are going abroad, or controlling your home environment or wi-fi.
Tapping into the potential
There is no question that AI has huge potential. What is not entirely clear is how this potential will play out, in telecoms or other sectors. The fact that ETSI’s members have thrown their weight behind network management systems suggests that improving efficiency and processes, and therefore increasing capacity and reducing costs, is uppermost in their minds. Customer services systems, however, are already with us, and it is surely only a matter of time before all telecoms providers are using them.
We hosted a panel discussion on Twitter where our telecommunications and machine learning experts explored these topics:
- What are the transformation opportunities facing the telecommunications landscape?
- How is #ML being deployed today?
- What are the similarities and differences between #ML for network optimisation and customer service?
- What key steps are early machine learning adopters taking to reap transformation benefits?
- Who do you consider to be innovators worth watching?
Highlights of this discussion can be found as a Storify.