Intelligent Customer Service with Machine Learning

SAP Machine Learning collects and analyses customer data. Based on an e-mail received from the customer and the language used in it, the customer's current basic mood can be predicted. A small smiley, which can reflect different facial expressions, helps the service agent to deal with urgent requests quickly. The confidence level at which a basic mood should be displayed can be set with a threshold in the machine learning scenarios.
In addition to the basic mood of the customer, SAP Machine Learning can extract further important information from the language used by the customer, such as order or product numbers, and make this information directly available in the ticket. Furthermore, it immediately recognises which national language the customer is using and can even translate directly if necessary.

An example of a practical machine learning feature of SAP Service Cloud is the recommendation of e-mail templates. To answer an incoming enquiry, a service agent usually first has to select an e-mail template from a pool of hundreds and then customise it in line with the customer's request. Machine learning helps the service agent with this process by immediately suggesting the three most likely and appropriate templates. This makes work easier for new service staff in particular and ensures a really easy start.

The machine learning of SAP Service Cloud adapts itself individually to each company. In order to initially train it for use, a minimum number of processed tickets must therefore already be available in the system, the data of which can be used for learning. Good results can only be achieved if the training data is of high quality and provides a good overview. Weekly data extractions from the ticket base continually feed and help improve the system so that it never stops learning, enabling it to deliver even more precise suggestions for all customer concerns over time.