5G is here. What does that mean for network operations?
5G and IoT platforms are growing more complex. As such, service providers are increasingly turning to cognitive technologies such as AI and robotic process automation in their telecom network operations. Is this hype or just plain common sense?
Running telecom network operations is more complex than it has ever been. It presents a whole catalogue of challenges such as legacy, multi-vendor, multi-RAN (2G, 3G, 4G, 5G), multi-technology (wireless, wireline etc.), as well as different billing and proprietary hardware-based solutions.
These operations can span across different organizational segments of any given Communication Service Provider (CSP), for example technical, marketing, customer care, product planning and strategy, and billing. With 5G and the much-anticipated explosion of IoT devices unfolding, CSPs are challenged with having to manage an ever-growing number of use cases, as well as new service requirements, virtualization and network slicing. At the same time, this presents an opportunity to transform network operations so that they are autonomous, proactive and contain minimal human intervention.
Our recent network AI & automation report highlights the growing importance of operations in order to provide a great customer experience. One of the overwhelming takeaways tells us that operations are not only central to cost efficiency but also have a significant impact on the customer experience. That’s why it’s even more important to modernize the way telecom network operations are being run.
As network complexities increase with 5G and IoT, it will be impossible to manage telecom network operations without the use of cognitive technologies.
What are cognitive technologies in the network?
For the uninitiated, cognitive technologies try to mimic functions of the human brain via natural language processing, data mining and pattern recognition. They are generally treated as a sub-branch of the broader Artificial Intelligence (AI) paradigm. In short, they are considered capable of performing tasks which are often performed by humans.
Introducing robotic process automation in telecom networks will make network operations autonomous, and thereby enable decisions which were earlier taken by humans. This can also reduce the Mean Time to Repair (MTTR). However, a “one size fits all” cognitive framework will not work for telecom operations. This is because the technologies that are used to capture information about the different environments are diverse, domain specific and use-case dependent. For example, natural language processing enables interaction with human users; robotic process automation performs repeated tasks that were earlier done by humans, active/passive probes and sensors captures network data, and an analytics system processes data to provide relevant insights.
As these systems continue to learn and evolve from historical patterns, they can easily identify potential bottlenecks and degradation in service.
The role of cognitive technologies in network operations
Telecom network operations involves huge amount of data processing and analysis, and are often characterized by highly manual, repetitive, time consuming and alarm-based actions. The adoption of cognitive technologies in telecom network operations provides plenty of direct/indirect advantages to the CSP. Below are just few examples:
- Content monetization - content can be monetized based on the customer’s usage behaviors and the most accessed media/video content in the cells/network. Instead of humans determining which content to move towards edge cloud, which to be cached and pushed to users for monetization and marketing aspects, cognitive network operations can take decisions in real time depending on the popularity of the content and it access rate.
- Customer segment insights - we have the concept of VIP customers, high ARPU customers, enterprise customers, and other segmentation in telecom network operations. Identifying such customer preferences, usage, behaviors, continuously learning and updating the previously learned preferences and providing them tailored packages can open a plethora of opportunities for the CSP. A combination of cognitive automation and AI enables these network insights.
- Intelligent customer care agents - cognitive intelligence can also be extended to the way automated chat bots and customer care agents respond to customer complaints. These agents can be advised in real time, learning in super quick time from thousands of other similar complaints, data points, billing plans, known and planned outages, customer’s history of communication with customer care etc. Based on analysis of multiple data points, opportunities for upsell or probability of churn could also be derived.
- Network Design and optimization - one area which has huge potential to be impacted by cognitive technologies are network design, optimization and planning. Network design and optimization are generally classified as a set of niche skills requiring years of experience to be called an expert in the area. Already, at Ericsson, we use AI to improve cell throughput gain, increase elastic RAN capacity, detect sleeping cells in the network, and predict power failures at cell sites. This is certainly one area, which shall see remarkable progress in the years to come with cognitive network operations.
- Reducing human errors - automation of daily network reports, routine tasks, anomalies detection and possible impact on business, identifying events leading to revenue loss. One of the major motivations towards automating these tasks is to reduce the risk of human error. Automating these tasks not only removes the possibility of human errors due to manual data analysis but also removes the false positives. Also, AI is more accurate than static dashboard and covers plenty of the metrices which are normally not seen in reactive operations. It not only improves quality of these tasks but also provides an option of moving those employees to other business-critical tasks.
The transformation from CSP to digital service provider
According to the Ericsson Mobility Report, 5G subscriptions are expected to shoot up to 2.6 billion amounting to 29 percent of all mobile subscriptions by 2025. Total IoT connections including wide area IoT, cellular IoT and short-range IoT will be more than a whopping 24.9 billion! By 2030, expected industry digitalization revenues for all ICT organizations globally is expected to be around USD 3.8 trillion. Up to 18 percent of this revenue, in other words around USD 700 billion is still up for grabs for CSPs.
CSPs must prepare early for this massive connectivity push in order to be a leader in enabling Industry 4.0. Here at Ericsson, we have been at the forefront of supporting this transformation journey, such as by enhancing network efficiencies by adding AI capabilities to baseband. Another example is our operations engine which brings together capabilities across network and IT operations, along with network design, planning and optimization, to improve customer experience.
Ultimately, 5G will transform communication service providers into digital service providers. To gain an edge, the time to begin that transformation is now.
Visit Ericsson Managed Services to find more insights about AI-enabled network operations.
I also recommend reading my other blog post where I investigate the 5G customer experience and the impact of AI and analytics. This area is also explored in more depth in our recent report Supercharging Customer Experience through AI and Automation.
Join our upcoming webinar
Networks are becoming rapidly more complex. In our upcoming webinar, Ericsson’s Head of Strategy for Digital Services and Head of AI & Automation for Managed Services explore how operators can apply AI to both optimize performance and lower complexity.
Join the webinar: Artificial intelligence enabling the 5G networks of tomorrow.