Data science and artificial intelligence are now ubiquitous factors in marketing. Despite or precisely because of their omnipresence, it is important to critically question the corresponding applications and not blindly trust the algorithms and methods offered. But many lack the time or scientific access to the new approach.

A concrete case study will show how a financial company’s data science team can support marketing and retain control over AI approaches. The goal is to develop a self-learning, automated process triggering the sending of e-mails and letters that best correspond to the expected customer interest at the time of shipment.

The system presented here includes interfaces for data import and export to fit seamlessly into the respective system landscape. The data is processed in an R or Python instance. A so-called self-learning algorithm is developed, which optimizes the calculation over time based on the customer reactions and the respective success of the individual measures. Possible pitfalls are identified and how to avoid them.

The case study presented has been developed for a bank, but also draws on experience in the insurance sector. The procedure can also be used analogously for existing customers in other industries (such as automotive, energy suppliers, telephone/internet providers, luxury goods, etc., where customer loyalty and up/cross selling play a major role.