White Paper – The role of Artificial Intelligence in network management

Cognitive computing has gained an admirable track record of systematic problem solving over the past decade. When IBM’s supercomputer Watson first appeared on the television quiz show Jeopardy! back in 2011, its ability to compete with (and outperform) human intellect was found most remarkable. Since then, Watson’s performance has increased by 2,400% and is today employed by over 350 Watson Ecosystem partners in a range of fields such as personality assessment, veterinary medicine and cyber security. The rise of these “virtual white-collar professionals” in business process optimisation opens up the question what role Machine Automation (MA) or Artificial Intelligence (AI), in its looser definition, will play in the future of network management.

A typical corporation today generates about 10x the volume of network events it did a decade ago. The growing complexity of corporate ICT ecosystems has increased the demand for intelligent coordination, configuration and collaboration between business units and suppliers. While the criticality of intelligent event management for ensuring resilient and stable cross-border services is of growing importance, customer expectations have similarly increased. The wave of AI and big data management opens up new opportunities for solving highly multifaceted problems faster, cheaper and more accurately than by human intervention. In particular, AI can optimise decision making among multiple network objectives and navigate across collective routes of cost, performance and user experience considerations. Greater volumes of quality information enhance the chances of getting it right. When you add logical inference and Bayesian probabilities to the equation the algorithmic advantage becomes unarguable. Consequently AI can increase reactiveness in situations of network incidents, while enabling both predictive and proactive problem solving through improved root cause analysis.

IPSoft is a company that is already bringing AI to the telecom arena of network management. IPSoft’s vision builds on integrating automatic and cognitive technology with IT and business processes, freeing up to 80% of currently allocated human capital resources to focus on value creation and innovation. The company calls its virtual engineer “Amelia 2.0”, which, similar to its predecessors, learns and improves through experience. Contrary to its earlier versions, however, it also possesses emotional responsiveness and contextual comprehension on a semantic level. Furthermore, Amelia’s architecture is designed to allow for scalability and to maintain resilience through peak times. In effect, Amelia acts as a virtual agent in providing technology helpdesk, contact centre, procurement processing and field engineer advisory services.

Likewise Blue Prism, a software company based in the North West of UK, offers to eliminate routine work and to enhance corporate agility through deploying robotic process automation for enterprise back-office administrative purposes. Blue Prism states its robot full-time equivalents (FTEs) cost about a third of current (human) offshore FTEs, leaning towards a future of decreased task-oriented offshore outsourcing. The automation benefits to the customer are realised through cost savings and resolution efficiency. The supplier benefits, however, are less straightforward as they face the conflict of interest balancing improved customer satisfaction on the one hand and potential loss of revenue on the other.

If cloud computing is considered the most disruptive force in IT outsourcing today, automation and AI might well represent the next big wave of change to the way business ICT systems operate. The reactive approach to network management is long out-dated and companies are looking for more intelligent, proactive solutions. That said, enterprise AI is no longer a conceptual idea of the future, but is already here and performing well. The growing sophistication of the software will improve the human-AI interaction and allow for greater outsourcing of ICT ecosystems to artificial information brokers.

Contract renewal presents a good opportunity for reviewing the technical, operational and commercial aspects associated with AI, while assuming the service within a currently outstanding agreement could appear more challenging. According to the McKinsey insight “Spend wisely, not more, on IT” published in January 2016, financial services’ targeted spend on automation, particularly in back-office application, can correlate up to 67% with business profitability. AI could come to flow through all layers of telecom systems, including business process outline, transmission, control, service and support. Ultimately it is up to the CIO/CTO to judge on the business benefits of AI and the extent to which this technology could play a key role in the future management of networks.