Airtel's data-driven transformation journey

Who: Airtel and Ericsson

What: Transformed network operations to a predictive, autonomous operations model

How: Deployed Ericsson Operations Engine, compliant with TM Forum’s Business Process Framework and Open APIs as the basis for a service-centric operations model focused on data-driven business outcomes

Results: Doubled the amount of automation in its network; 69% of alarms fully automated; reduced mean time to repair (MTTR) by 29%; reduced network unavailability by 47%; 26% improvement in customer experience

India is famously one of the most hypercompetitive mobile markets on Earth – and one of the most challenging. Competing on price makes little strategic sense in a market where ARPU is around $2.00 a month. Indeed, with churn rates in India reaching 2 %, competitive differentiation really comes down to customer experience and customer care. For Airtel, the two go hand in hand – tracking customer complaints serves as a powerful metric to track network performance.

However, Airtel found that its traditional, reactive approach to network operations was people and OpEx-heavy and not up to the task of maintaining and improving the customer expectations of its 340 million subscribers. Together with Ericsson, Airtel embarked on a transformation journey to shift its network operations to a predictive, autonomous operations model.

The results of that shift on Airtel’s service uptime and service fulfilment metrics have been dramatic, to say the least. Airtel says it has doubled the amount of automation in its network and has reduced work orders per nod, MTTR and network unavailability – all of which has produced a considerable and measurable improvement in customer experience.

Addressing the challenges

When planning their automation journey with Ericsson, Airtel pinned down three key challenges it would have to overcome, starting with the sheer scale of the project, as well as the heterogenous nature of the network itself.

“Airtel is a multivendor 2G/4G network environment carrying 112 petabytes of traffic per day,” explains Randeep Sekhon, CTO of Airtel. “We have 220,000 towers across India with typically 12 radios per base site.”

Other key challenges included the complexity of legacy processes, tools and platforms to be transformed, and upskilling the field teams that are installing up to 25,000 new base sites every year.

“Airtel needs a skilled workforce across large geographical areas, including remote regions, to operate our growing network,” says Sekhon. “Upskilling field teams in remote locations proved challenging and resulted in high attrition rates.”

Airtel tackled these challenges with the Ericsson Operations Engine (EOE), a data-driven technology-agnostic operations solution that supports multi-vendor environments and leverages AI, automation and data analytics-based capabilities to handle the growing complexity and scale of Airtel’s network operations.

TM Forum assets help deliver service-centric operations model

Airtel’s Ericsson Operations Engine implementation relies on a multi-layered process framework based on TM Forum’s Business Process Framework (eTOM).

Conceptually, Ericsson Operations Engine has three major value chain steps – Create, Sustain and Evolve which align with the Business Process Framework’s Strategy and Readiness, Operations and Enterprise Management and Common Function layers. It covers a total of 26 processes designed to support these value chains with consistent and high-quality, end-to-end process flows. Ericsson Operations Engine’s alignment with the Business Process Framework provides the flexibility to configure and customize tool functionality out of the box.

At the core of the Ericsson Operations Engine is a portfolio of APIs, including TM Forum Open APIs and APIs compliant with REST and TOSCA to better integrate with customer networks, as well as within the Ericsson Operations Engine itself to support execution of data-driven processes. Ericsson Operations Engine also aligns with the six dimensions of the TM Forum Digital Maturity Model (DMM), and references best practices from the TM Forum Open Digital Architecture (ODA).

With the Ericsson Operations Engine as the engine of Airtel’s network operations initiative, Airtel developed a new service-centric operations model with end-to-end capabilities and best-in-class tools and processes that leverage AI, automation and data insights to deliver optimized business outcomes.

What data-driven network ops looks like

Airtel is delighted with the results, citing a long list of tangible benefits from its journey to become an automated, data-driven operation.

For example, in terms of efficiency, says Bradley Mead, Head of Network Managed Services of Ericsson, “Airtel has implemented more than 300 automation rules and AI use cases to improve service quality that are already providing correlations, which otherwise would not have been seen.”

Airtel has more than doubled the amount of automation in its network, with 69% of all alarms automatically correlated and resolved without human intervention. Another 14% are semi-automated, while the rest still require a human to manage the process.

Machine learning plays a key role here by helping apply data correlation to alarms, says Mead: “Based on applied analytics, Airtel was able to identify and isolate 789 alarms that were sent to field management that can now be solved with automation by the network operation center, resulting in 15% less work orders per node.”

Also, by using analytics and automation, 50% of work orders have been enriched by automated root cause analysis. Root cause analysis and real-time access to alarms has also helped reduce field visits.

On the network performance front, Airtel says it has reduced MTTR by 29% and network unavailability by 47%. “This is a paradigm shift in that Airtel and Ericsson have changed the way we approached the uptime concept, shifting from measuring network uptime to measuring service uptime,” Sekhon says.

The analytics and automation built into operations has also enabled Airtel to keep up with the needed pace of continuously growing its network to stay competitive – according to Bradley Mead, 3.1 million changes happen in the network monthly with 99.6% accuracy.

All up, the end-to-end view of service operations across network, IT and cloud have boosted Airtel’s service uptime and service fulfilment metrics that it credits with a 26% improvement in customer experience. And that’s down to Airtel’s ability to use data-driven, predictive operations, to resolve potential network issues before the customer has a chance to notice that anything is wrong, says Airtel’s CTO.

The transformation journey continues for Airtel, with strategic ambitions towards a fully automated, predictive operations. As Randeep Sekhon says, “Predictive is where we really want to be paired with closed loop automation– then you’re really hands off.”