Service assurance is officially graduating from the days of dashboards, tickets, and engineers scrambling to find the source of problems for rapid root cause analysis and proactive remediation.
As AI moves deeper into the network stack, there has been a great deal of experimentation to find out how to best tune the network using AI.
“Today’s networks are 150 times more complex than traditional networks, and the only way to address or manage this operational complexity is through continuous testing and complete automation,” said Anil Kollipara, vice president of product management at Spirent, in a recent presentation.
Over the past few months, a clear trend has emerged. Solution providers are incorporating AI into their portfolios to achieve greater levels of autonomy, observability, and speed of resolution. The goal is to make service assurance low-touch for operators, and for many operators, complete automation of the service assurance process remains an immediate goal.
This change was a long time coming. Network operations have had a bad reputation for quite some time. It is seen by insiders as a thankless job, with long hours, tedious work, and blame when things go wrong.
With the responsibility for network testing and service assurance now shifting from equipment vendors to service providers, there is a natural imperative to find ways to improve service quality control and reduce repair times.
There is evidence of an increasing degree of autonomy among operators in service guarantees. According to the GSMA Intelligence report, three-quarters of surveyed operators are automating their service assurance processes, with more than a third saying the majority of their processes are already automated.
AI may not be evaluating everything yet, but AI-driven service assurance is definitely gaining traction among carriers. Importantly in all three areas, the role of AI is becoming increasingly important across domains.
Root cause analysis
“The process of getting to the root of a problem, or root cause analysis (RCA), is a very laborious and tedious process, even with automated cycles in place,” Kollipara said.
RCA has multiple steps including, but not limited to, defining the problem, collecting artifacts, performing analysis, diagnosing, and identifying root causes.
— That’s challenging.
AI provides some very specialized capabilities that reduce this weeks-long process to minutes. For example, you can scan large data sets almost instantly, identify patterns within the data set, and create automatic correlations across the system.
This makes connecting the dots, which is essentially the work of root cause analysis, much easier and more reliably automated. Within minutes, AI examines thousands of data points from network logs, telemetry, and KPIs to uncover where incidents occur and why.
Currently, RCA is one of the top AI use cases in communication networks, according to several studies.
Proactive anomaly detection
AI workloads are, for lack of a better word, chaotic and subject to frequent anomalies and deviations.
AI models offer a unique opportunity to solve them. A good AI model can identify unusual patterns and outliers in large datasets with 100% accuracy. This is a great way to capture deviations in network performance.
While AI continues to make networks highly complex, it can help providers cut through the noise and proactively detect issues to reduce outages.
Level 4 and Level 5 autonomy is an ambition for most operators, and AI-driven proactive anomaly detection is considered one of the fastest ways to achieve it.
customer analysis
AI-driven analytics is also one of the most practical AI use cases in service assurance. AI models are good at reading poor user experience, usage patterns, upsells, and other analytics that can indicate churn. This allows us to predict the risk of customer loss and
According to the GSMA report, the majority of carriers are already using AI for customer analytics, with 80% using AI to generate customer-related insights and 63% using AI for customer complaint analysis. Additionally, 34% said 51% to 75% of their analytics processes today are AI-driven.
