Predictive system | Mobi724

Personalization of the offerings associated with payment cards with the help of reinforcement learning

Mobi724 offers services to companies in the financial sector, mainly banks, by sending promotional offers associated with credit cards.

To do so, Mobi724 uses a prediction system associated with customers’ payment cards. The tool, based on reinforcement learning algorithms, was developed in partnership with CRIM. Using it has made it possible to more accurately model consumer habits. The goal of this system? Targeting the preferences of millions of users in order to provide them with adequate offers.

A FinTech company, Mobi724 offers value-added services to banks and financial institutions. The core of its activity is data analysis.

Mobi724 is working in collaboration with CRIM to define a relevant analysis model. The collaboration was initiated through the “Innovation” program set up by Investissement Québec. CRIM’s expertise in data science has made it possible to more accurately model the consumption habits of Mobi724’s clients’ users.

The idea is that if the tests are satisfactory, the technology developed will be transferred to FinTech’s company. “The key word is appropriation. We will make sure that Mobi724’s employees understand the technology and methods well enough to continue and repeat new development iterations,” explains Michel Savard, Senior Data Scientist at CRIM.