The Challenge:

A major Canadian bank wanted to use artificial intelligence technology to improve its call center operations. The bank wanted a system that would help them identify which credit card customers they should be prioritizing during the collections process, when to proactively reach out to them and also how to personalize communications in ways that resonated with individual customers. The bank already knew who would repay old debts and who would react positively to the bank contacting them, but that was only within a small subset of its customer base.

The Solution:

Dessa’s business and technical teams collaborated with the bank to build an AI system that could take historical customer behavioral data and predict a customer’s reaction when reached out to a call center operator. The bank’s previous model would assign a “risk score” to determine whether it should call the customer, but it didn’t take into account the probability of whether any communication would influence a customer’s decision. An AI model took previously known data and inferred a risk score for each customer based on similarities of their banking behaviour, regardless if they were contacted or not.

Results:

Dessa completed its AI model over a three-month period and generated significant results. The team used an AI algorithm to perfect its model using an academic research paper that came out during the project. The Dessa model helped the bank’s call center to identify whether a customer should be contacted by 20% compared to the previous system, ensuring better management of existing resources and a boost to productivity.