Home / Bank Support Agent
Case Study

Bank Support Agent

Rainbird was integrated with IBM Watson’s Natural Language Processing (NLP) front-end to deliver for a bank a chatbot that can resolve over one hundred different trajectories of enquiry.

Rainbird was selected to work with the retail arm of a multinational banking and financial services company. The bank operates an extensive branch network and uses a contact centre to support their millions of customers in their banking needs.

The problem

The bank receives thousands of enquiries every day, many of which relate to regular and recurring payments such as direct debits and standing orders. Staff have access to vast amounts of information internally which they rely on to resolve these queries.

Like many such information stores, the content is very substantial and wide-ranging. Staff have to navigate these articles and on finding an article that might be relevant, must scan through pages of detailed information to find the section relevant to the specific query they are dealing with.

The articles contain general guidance, and the employee needs to determine how best to put this in context for each individual customer. This process can be time consuming, for both staff and customers. Last year, employee turnover in banking was reported to be at a ten-year high. Banks can cover this loss of experience and know-how by encoding their expertise into a knowledge map.

Knowledge degradation is also an issue, in areas of the bank where staff turnover is higher, such as in contact centres. Expert knowledge can be lost if experienced staff leave, and the consistency of judgements can suffer as a result.

Less experienced staff tend to seek out experts, who then become a focal point for queries, detracting from their own day-to-day responsibilities.

The solution

The bank engaged Rainbird to build a tool that any employee could use to answer client questions and resolve problems with complex recurring payments. Rainbird was integrated with IBM Watson’s Natural Language Processing (NLP) front-end to deliver a chatbot that can resolve over one hundred different trajectories of enquiry. IBM Watson is used to identify the intention of the employee asking the question, and Rainbird then takes over to reason over the decision using its Decision Intelligence technology, handling the staff consultation and leading to a contextual solution. The result of this collaboration was a Rainbird knowledge graph incorporating all of their knowledge with the information from the internal knowledge management platform.

The resulting chatbot is available to staff who can efficiently consult with it to rapidly solve customer queries. The result includes a high-quality contextual recommendation including detailed advice on how to meet the customer’s needs and signposting to the actions required.

Take your first step towards transforming your organisation with Decision Intelligence, powered by explainable AI.

If you want to find out more about how Rainbird can help you, contact us to talk to an expert.