Transcom has been delivering a debt collection service, both inbound and outbound, to a large European telecommunications company for years. With the new digitization processes of the telco industry, the client required a new debt collection management approach in order to make operations more efficient and reduce costs.
After a detailed analysis of the debt collection processes, we designed and deployed a machine learning solution to meet the new service challenges.
Machine learning is a subfield of artificial intelligence (AI) that focuses on developing algorithms and models capable of automatically learning patterns and making predictions or taking actions based on data, without being explicitly programmed. It involves training a computer system using large amounts of data and allowing it to automatically improve its performance over time through experience, rather than relying on explicit instructions.
In the context of debt collection, we used machine learning to analyze customer behavior and payment patterns, identify risk factors, and optimize the client’s strategy to increase the efficiency and effectiveness of debt collection processes.
For the realization of the project, more than one million contacts were used to obtain the best algorithms applicable to the applied machine learning solution.
The solution is based on a daily loading of customers into the system with updated debt. These customers are characterized by means of algorithms to predict whether or not they will pay their debt.
Depending on the probability of payment, those customers with the highest probability are automatically sent through an IVR system to our client's payment gateway to make the self-payment and thus reduce agent operating hours.
The rest of the customers with a lower probability of payment are given different collection strategies, from automatically sending the call to high-performance agents to proactively making outbound calls.
This machine learning solution has allowed not only to improve collection ratios but also to reduce agent operating hours. This way, we were able to increase self-pay by 47% and reduce operating hours by 1.000 hours per month. As a result, the client’s operational costs were significantly reduced.