CompuScore Launches Alternative Credit Scoring Engine Using Telco and Utility Data for Unbanked South Africans

JOHANNESBURG – CompuScore, the credit bureau formerly known as Compuscan and CompuCredit, has launched a new alternative credit scoring engine that leverages mobile network operator data, utility payment histories, and transactional behaviour to generate credit profiles for an estimated 11 million unbanked South Africans, the company announced Tuesday.

The rebranded entity introduced "CompuScore Alternative," which integrates real-time API connections with major telecommunications providers and municipal utility systems. Unlike traditional scores reliant on bank account activity, the system analyses prepaid airtime top-up consistency, utility bill settlement timeliness, and informal retail transaction behaviour. The machine learning engine, trained on over 2.5 million anonymised consumer records, predicts creditworthiness with reported accuracy rates of 78 percent for first-time borrowers. Chief executive officer Michael Povey stated that the rebranding to CompuScore reflects the company's strategic pivot from legacy credit reporting to dynamic, alternative data-driven assessments, positioning the engine as a gateway for the "credit invisible" population.

The launch follows the National Credit Regulator's 2025 directive encouraging alternative data adoption. Major lenders, including Capitec Bank, have signed pilot agreements to test the engine on entry-level credit products. The engine offers risk-pricing modules with interest rates between 19 percent and 32 percent, significantly lower than the 60 percent rates typical of unsecured micro-lenders targeting this segment.

However, data privacy advocates have raised concerns regarding the use of telecommunications data, which falls outside the traditional National Credit Act framework. Lawyers from the Centre for Applied Legal Studies question whether consumer consent for such granular data aggregation is fully informed under current regulations.

CompuScore has responded by implementing a dual-consent workflow, requiring explicit opt-in authorisation separate from standard network operator terms of service. The company has submitted its algorithmic model for independent auditing by the University of the Witwatersrand.