Optimizing the complex world of customer service assurance in 5G

Emerging technologies like AI, machine learning and predictive analytics will bring in a shift from reactive to proactive service assurance to keep customers happy

One of the biggest challenges for communication service providers (CSPs) is to maintain consistent quality of service and keep customers happy. This is not an easy task when the customer base in question is as large as 1,147 million. Telecom infrastructure that is required to service this customer base is equally large and as networks become more complex, problems become harder to identify and resolve. 5G, being a new technology, will add another layer of complexity in network management and in the initial phases of its roll-out, delivering an optimal quality of service to customers will be a big plus for the CSPs.

It is no longer an option to manually investigate network faults before they show significant impact on customer experience. There is only so much that human resources can achieve, especially when overwhelmed by an abundance of alarms, alerts, and their subsequent trouble tickets when each new issue occurs.

As degrading customer experience continues to haunt CSPs, the Telecom Regulatory Authority of India (TRAI) maintains a close watch on their performance through quarterly performance monitoring reports. For instance, to monitor the call drop rate in a network, TRAI uses two parameters—DCR spatial distribution measure and DCR temporal distribution measure.

Therefore, to manage network performance issues and faults, operators need new, more efficient, and less time-consuming solutions. The traditional approaches to service assurance are fast becoming redundant as technology infrastructures and operations go through a period of transformation with constant network innovation. This brings forth the case for automation in network fault management, driven by AI, machine learning, and analytics.

The traditional approach is no longer relevant

The latest generation of platforms will need to consider increasing interdependence across various domains and service offerings within its more dynamic technology environment, increased data volumes, and the acceleration towards autonomous operations.

Trouble ticketing is a key function that is used in support of modern service assurance. This covers receipt, assessment, correlation, and resolution of incidents that impact the network. This sets in motion both reactive and proactive processes, which if executed correctly, help maintain or increase the network quality of service for customers. Optimization of this process requires complete visibility in the network environment, both horizontally and vertically. It also requires a timely and careful approach to decision-making to optimize process efficiency.

Traditionally, operators have had to work with distinct service assurance platforms to serve varying domains. These legacy systems have lost their relevance in today’s complex and dynamic ecosystems of multiple in-house supplier and partner solutions. Conventional integration methods are expensive and inefficient, and customization is passéwhen it comes to delivering the necessary levels of efficiency and quality expected in today’s complex networks.

Top this with the complexity of 5G networks and traditional service assurance is no longer viable.

AI and machine learning take service assurance to a new level

The need of the time is for the CSPs to react quickly to evolving customer needs and new business opportunities. Hence, the gap between service assurance requirements and capabilities has to reduce. The operators are therefore making a move towards more proactive, predictive, and autonomous operations. The objective is to have a system where decision making can be entirely automated, without the need of human resources devoted to repetitive, prosaic tasks and focus on critical customer impacting issues.

This is where AI and machine learning move in. These new technologies can facilitate closed-loop automation to deliver new applications and services quickly. The result is network issues can be resolved within a fraction of the time previously required.Another big shift that AI and machine learning will bring in is the move from reactive service assurance processes that fix existing problems, to a predictive approach that pre-empts the emergence of problems before they can impact the network.

The transformative power of AI will help address common underlying problems across multiple systems. Operators can use big data, machine learning and analytics to detect patterns in monitoring, capacity, and automation data across complex technology infrastructure, thus improving their overall service assurance processes. Next level will bring in closed loop automation to minimize human involvement and head towards zero-touch.

Predicative analytics and the shift toward AIOps

Predictive analytics will play a significant role in achieving zero-touch closed-loop automation with service assurance. The ability to predict network faults and failures before they even happen will transform how networks are managed. While the application of such capabilities to real-world scenarios is not new, CSPs still need to evolve their processes to support the outcomes.

This is how it works. First, establish a consensus among internal teams as to what should be predicted. Once it is decided what qualifies as a high priority issue, it can be immediately addressed. Note here that identifying and prioritizing high-priority issues will be part of the learning curve. In order to address these AI-driven predictions, CSPs will have to adapt their existing processes and competencies and ensure their operations are more flexible and open to rapid change.

The broader implication of this is the adoption of AIOps by the industry, in general. CSPs are already experimenting with these new concepts in service assurance. For example, they are applying unsupervised machine learning techniques to historical bodies of data held by them. Thus, an analysis of equipment alarm and failure records will lead to an understanding of incident patterns and its likely recurrence, allowing the definition and implementation of preventive maintenance processes.Similarly, consider analysis of historical ticket data in terms of a current issue being investigated vis-à-vis the resolution that fixed the issue. This can help generate a knowledge base that powers a “next-best-action” recommendation engine.

These applications of AI can help CSPs manage their networks with greater efficiency, by cutting the human time spent in analyzing the problems and instead allowing them to focus on customer experience impacting issues.

Innovation is the way to successful service assurance

In a dynamic and rapidly evolving telecommunications landscape, service quality is key. Customer expectations are only set to rise with the introduction of 5G applications and services. Tolerance for disappointing performance, unreliable availability and slow responses to problems will further decline. As the velocity of network provisioning increases through automation, CSPs will have to ensure that their service assurance tools and processes can keep pace with these advances and become more automated as well.

Emerging technologies including AI/ML hold the key to improved customer experience. By rapidly applying innovative approaches to operations process, CSPs can ensure the services consumed today and in the future keep running, even in the face of growing network complexity.Amidst severe competition as 5G services go live in the country,it becomes all the more important for a CSP to keep its customers happy and respond to new business opportunities in real time. Automation of service assurance will hold the key.

Authored by: Jugad Bawa, Director, Sales, South Asia and Middle East, TEOCO,India