As a legal protection insurance company for individuals as well as private companies, DAS receives a large number of invoices from bailiffs, lawyers, experts, etc ... These service providers, who generally do not have fixed fees, send their invoices to DAS. Of course, DAS is aware that some of these experts have a tendency to bill too much for the work that has been executed. However, going through all the bills to check if the invoiced amount is acceptable is impossible:
Due to the volume of these interventions, an in-depth analysis of these invoices costs more than the amount that can be disputed. 

This is why DAS asked Sagacify to develop an automated solution that determines whether the invoice amounts are abnormally high for the task, and whether disputing an invoice is necessary.

‍

The model we developed collects information about the claim in different databases. Data such as the beneficiary, the parties involved, the content of the invoice, but also history of billing expert, and the same type of cases across different experts. With this information, the model is able to predict whether or not the invoice should be disputed with a high precision.

Key elements:

• Use of highly sensitive data. 
• Need to understand the business process to know which features are relevant. 
• Need for close collaboration between Sagacify’s technology staff and DAS’ business staff to assess whether the predictions of invoice dispute are justified.

‍