How do you prevent a service desk from spending a lot of time answering the same questions over and over again? The Rabobank designed chatbot Billy to handle recurring questions about bills. This digital source of information uses SAP Conversational AI (CAI), SAP ‘s chatbot platform for automating customer contact.
Rabobank is keen to automate repetitive actions. Automation ensures that, for example, daily checks are carried out more efficiently and faster, and also continue at night. Because repetitive actions are eliminated, employee satisfaction increases in many cases.
The financial services provider eliminates manual actions, among other things with the help of robotic process automation (RPA). For example, a robot checks daily whether the money for, for example, a mortgage or insurance has entered an account, checks whether the description states where the money should be sent, and, if necessary, carries out the transfer.
“In the daily operation we have to deal with a lot of manual work that comes back every day. We want to get rid of that in order to relieve the employees of Rabobank as much as possible ”, says Robert van Uden, Senior Solution Engineer at Rabobank. “We do this by using RPA, but also chatbots.”
Rabobank built the chatbot Billy to handle invoice-related questions from suppliers. “Colleagues spent a great deal of time answering the same questions over and over, such as ‘when will my bill be paid?’. The chatbot was able to take over some of the work there. ”
Rabobank selected SAP Conversational AI as the platform for chatbot Billy, because it connects with existing systems, such as SAP ECC.
Security and privacy guaranteed
Rabobank itself took care of the implementation of the chatbot, with the assistance of SAP. This assistance was particularly useful in the field of security. “You are dealing with a cloud chatbot that will collect data internally from Rabobank. Then you have to be sure that the privacy and security of that data is guaranteed at all times, ”explains Van Uden.
It was also important for the success of the chatbot that the scope of the questions was limited as much as possible. Robbert-Jan Wanna, Business analyst Robotics & Artificial Intelligence at Rabobank: “The broader the scope of the questions, the less efficient the chatbot becomes. We therefore asked the department to make an inventory of the questions that often come in. After cleaning, we clustered the questions into only four categories, including an ‘other’ category, and determined the responses per category. That way the answers are always correct, and you ensure that the more complex questions reach the employees. ”
Billy went live when the Netherlands was under the spell of COVID-19 and went into a lockdown. According to Van Uden this was perfect timing afterwards. “All colleagues suddenly had to work from home, and had to deal with changes. Billy made sure to relieve this team. The employees no longer had to look up the standard invoices and had enough time for the specific questions that were abundant during that period. Not only colleagues but also suppliers are positive about the contact with Billy “.
In this 24-hour economy, suppliers also have questions outside of office hours. The chatbot can therefore give them direct answers to concrete invoice questions at three in the morning.