Our client is a major Australian utility company. At the time of engagement, their CRM system contained over three million records, many of which were duplicates and represented the same customer despite differences in the information contained in the individual records due to spelling errors, changes of address and others. The company had hundreds of thousands of unpaid bills representing tens of millions of dollars in debtors. They needed a solution to clean up the CRM in order to accurately pursue the collection of unpaid debts and streamline administrative processes.
PAG developed a set of data matching algorithms apply advanced natural language processing (NLP) and other techniques. The fuzzy logic based matching algorithms gave significant improvement in performance compared to traditional methods. Machine learning is also used to optimise match acceptance thresholds, allowing the system to move toward a complete automation and away from any manual review processes.
The PAG solution substantially increased the client’s confidence in its administrative processes. It has allowed the company to maintain a clean and accurate CRM database going forward, and has allowed the organisation to follow up millions of dollars in outstanding invoices with confidence.