The pandemic made Türk Telekom more acutely aware of the importance of customer experience and the consequences of problems that affect services. Turkey’s biggest operator also recognized that for many years, key performance indicators (KPIs) for customer experience were on the business side, rather than the network, although it is the most fundamental factor in customer experience. The operator decided to steer its AI in Operations project towards customer experience as it had already proved itself in making network operations more efficient, since its inception in 2019. Now the project is interpreting network KPIs to evaluate problems before customers have reported them, or are even aware of them, to improve service quality. It is working to detect the real cause of customers’ complaints, for example, by assessing the performance of equipment on their premises. The operator also wants to find actionable solutions to different problems for different teams.

Outcomes

11 million

DSL lines analyzed weekly


40,955

services analyzed for anomalies and saturation


533 CPE

replacements analyzed for a 16 day period

in the last quarter of 2021

TRY 2.5 million

revenue increase gained

($144,271)

Deployed 500 cloud native pods for core network and CPE capacity at

100,000

sites across four zones in India

Deployed 500 cloud native pods for core network and CPE capacity at

100,000

sites across four zones in India

Deployed 500 cloud native pods for core network and CPE capacity at

100,000

sites across four zones in India

Deployed 500 cloud native pods for core network and CPE capacity at

100,000

sites across four zones in India

Yusuf Kıraç

CTO

“Turk Telekom operates a vast network of fix, mobile and broadband technologies, serving more than 52 million subscribers. A better customer experience is one of the highest priorities in a competitive environment, and our network operations are based on customer experience to provide such level of competency, says Yusuf Kıraç, CTO, Türk Telekom. “To proactively manage our network, we employed AI and machine learning methods in addition to conventional ones. Our project relies on anomaly detection and predictive maintenance that is customized for network operations to solve our customer-affecting problems.”

  He adds, “Studying network parameters and KPIs, alongside those generated by the business support systems with the use of flexible open-source technologies and robust methods based on TM Forum frameworks, provided a successful result.”

Applying the power of AI

Türk Telekom sought to address all these matters by applying AI and machine learning to data analysis to generate automated fixes. AI in Operations has a more than 80 integration points throughout Türk Telekom’s data infrastructure, on the business and network sides, so different types of data can be correlated for complex analyses.

Using this multi-domain data, the operator adopted three approaches: Anomaly and saturation detection for MPLS services which provided the opportunity to upsell; anomaly detection in xDSL broadband services to improve customer satisfaction by automatically optimizing service parameters; and identifying instances of when customers’ home equipment has been replaced unnecessarily and other erroneous decisions by the field team that led to repeated complaints.

TM Forum assets used

Find out more about how TM Forum membership can help your business


By signing up for this case study, you agree to the Privacy Policy