How to drive autonomous operations maturity with AI

Commercial context

CSPs are essential to digital transformation across almost all industries – but to meet that challenge, individual CSPs must themselves achieve digital transformation of network operational models. This means wholesale adoption autonomous operations, in part to enable the personalized customer experiences their vast new range of potential customers expect.

This is no mean feat – networks are only becoming more complex in the age of 5G, making cross-domain demarcation more difficult, and fault-handling experience summarization depends on highly specialized expert interventions into heavy workloads. Such operating expenses must nonetheless be limited to ensure profitability in this new complex commercial context, and dissatisfied customers are unlikely to wait for long before contributing to churn count.

Solution

To help address this complex set of challenges, the Intelligence and value driven digital operation transformation for network Catalyst is creating an intelligent operations framework based on AI to develop a closed-loop automation solution for autonomous operations scenarios, enabling CSPs to identify customer demand and potential revenue. This will be achieved by introducing an architecture framework and intelligent operations foundation based on TM Forum assets (including GB1042 and GB1042A on AOMM, TR284 AI Closed Loop Automation Implementation Architectures, and IG1292 Value Operation Framework) that can be built to phase in other core processes over time.

The primary goal of this solution is to demonstrate how AI technologies can be applied in the changing network environment by correlating and calculating massive quantities of cross-domain data, to then quantitatively analyze the impact of network problems on quality of user service. Network adaptability and governance can then be equipped with self-awareness, self-analysis, self-decision, and self-execution capabilities with minimum human intervention.

Applications and wider value

CSPs can therefore drive operational efficiency through MTT-resolution to minimize service impact and reduce customer churn - achieving over 90% accuracy in both network demarcation and root cause analysis via spatial-temporal alarm mining, and a clustering algorithm operating on the knowledge graph-modelled network topology with fault propagation. This is well beyond the capability of traditional human-assisted correlation – and the system can furthermore proactively explore undetectable service impact issues through a graph neural networks model, reducing data traffic by over 14%.

The business value of this project can be shown in satisfaction of the ‘R.I.S.E’ values expounded in the Value Operation Framework (VOF) of IG1269 and IG1292:

1. Revenue – by helping reduce customer churn and identify areas where customer demand is not met by existing network resources
2. Innovation – by helping CSPs achieve maturity in digital transformation through autonomous operations
3. Satisfaction – by delivering zero-trouble experiences for CSP customers
4. Efficiency – by driving operational efficiency through automation and greater accuracy in fault detection and resolution

“The most important benefit of the Catalyst is that we have validated the values of leveraging AI technologies to drive our digital operation transformation, and the solution is highly scalable as it shows consistent business outcomes across CSPs,” explains I Gede Darmayusa, Director & CTO at PT XL Axiata. To find out more join us at DTW Ignite in Copenhagen between 19-21 September 2023, or get in touch with the Catalyst team directly here.