Our Solutions / Services

To detect fraud, our team follows the following methodology.

An initial virtual meeting is held to define the tasks that we are going to carry out and define the items that are required from the client. 

Then our Artificial Intelligence team together with experts in the business including the users of the company analyze all the data collected and extract the essential characteristics (call destination, amount of data traffic, location, identity, orders, SIM card changes, payment method, etc. ). 

A first chunk of test data is received, analyzed and cleaned, leaving only the main fields that correlates with the events we would like to identify. 

Different models of IA are tested and evaluated to identify which performs better for predicting fraud events. 

Finally the model is tested with large volumes and deployed into production the model with the code incharge of getting the data, evaluating each event through the model , generating alarms if required.