• Topics
  • Research & Analysis
  • Features & Opinion
  • Webinars & Podcasts
  • Videos
  • Dtw

How data and machine learning can proactively increase revenues and customer satisfaction

Avoiding service errors is essential to reducing churn – and for CSPS that means using machine learning to understand customers through verbal and textual information

Alasdair Riggs
28 Nov 2022
How data and machine learning can proactively increase revenues and customer satisfaction

How data and machine learning can proactively increase revenues and customer satisfaction

In the competitive market of telecommunications, the ability to identify network malfunctions, billing errors and fraud before they impact customer experience is crucial – many CSPs however still tend to manage business assurance issues manually, as isolated events, despite their potentially serious implications for revenue and churn. And, as CSPs roll out multifaceted 5G services in the months and years ahead, manual customer feedback analysis is increasingly unlikely to help them keep up with service enablement and performance issues before customers lose their patience.

The Business assurance - listening to your customers’ voice Catalyst seeks to address this challenge by using the power of machine learning technologies – such as voice-to-text transcription, and sentiment and text analysis – to analyze customer interactions. This can identify commonalities among service issues, thereby enabling proactive approaches to business assurance.

Three CSPs (Batelco, Orange and Telefonica) are participating in the Catalyst, which also involves vendors Amdocs, Solvatio and Mobileum. “We use multiple advanced machine learning algorithms to ‘understand’ information from customers’ verbal, textual, and contextual interactions with the CSP,” explains Hernán Delgado of Telefonica Argentina. Identifying commonalities enables the CSP “not just to solve issues affecting complaining customers, but also proactively for other customers who might be affected by the same issues.”

Catalyst participants say trials across various use cases have demonstrated the value of machine learning in reducing the time it takes to detect and resolve billing errors, service issues, network malfunctions, fraud incidents and ultimately customer complaints. They believe this approach will deliver a significant increase in customer satisfaction, resulting in lower churn, and fewer and shorter calls to contact centers, thereby cutting costs and boosting revenues. “Using this innovative methodology, we can be proactive and rehabilitate clients who have had a problem in their allocation of payments,” adds Delgado. “Additionally, ‘listening to the voice of the customer’ should detect other problems in a company's processes.”

One trial for instance used the solution to identify cases where online gamers weren’t receiving the promised quality of service from a 5G network. It used artificial intelligence to analyze customer sentiment/feedback analysis and then automatically adjust the assurance functions in the network. The solution was able to accurately identify customers who hadn’t complained, but were affected by the negative experience. The CSP was then able to systematically improve its products, processes and controls.

In another trial, the solution was used to resolve an issue where a CSP was receiving complaints from customers unable to make or receive calls because their SIM cards had been deactivated and replaced by unauthorized third parties. Using machine learning to look at the complaints, the solution detected that many of these were related to a SIM swap and that there had been an increase in these complaints compared to previous weeks. The CSP was then able to warn other customers whose SIMs were swapped to verify that this was not the result of fraudulent activities. The solution was also able to establish that the fraudulent swap was performed at night-time by people from a certain ZIP code, which allowed better targeting of interventions.

Zahra Eid of Batelco, a leading CSP in Bahrain, anticipates the solution will enable proactive and simultaneous resolution of many customers’ issues and reduce calls to its call center. The CSP also expects to benefit from a ‘360-degree view’ of its customers and full customer journey-mapping analytics. This improved knowledge of the customer allows CSPs to detect, correct and prevent incorrect dunning processes and billing errors, saving time, resources and customer goodwill.

Arnold Buddenberg of Orange says the ability to capture customer feedback from different touchpoints is essential to identifying the “behaviour-knowledge patterns” necessary for proactive customer experience management. By using the solution to optimize customer engagements, Buddenberg expects to “reduce revenue loss and churn, improve renewal campaigns, and also increase both revenue, and appreciated trust.” As networks and connected services become ever more sophisticated and demanding – while consumer patience for error and delay continues to diminish – this innovation will surely be of interest across the industry.