Machine learning with SAP Service Cloud

7 June 2018

Uphaar Agrawalla

Uphaar Agrawalla

Solution Architect - SAP C/4HANA

You’ve probably heard the buzz around automation and machine learning and wondered what it's all about and where it fits within your customer engagement strategy. I'll try to explain by taking a customer service journey as an example. But before we do that, let me share...

...a personal experience

Not so long ago, the camera on my phone started malfunctioning. The phone was still under warranty and online research showed that it was a common issue that required sending the phone back to the service centre for repair as it was a manufacturing defect with the camera.

With this knowledge, I contacted the customer service team to report the issue, with the expectation that they will tell me that this is a known issue and the phone will be repaired free of charge.
Instead, I was told that the warranty coverage could not be confirmed until a technician has looked at the problem and that there was still a chance that I could be charged for the repair if the fault was not covered under warranty. At this point I felt:
  • That I knew more than the agent after a little online research.
  • Frustrated, as I couldn’t get confirmation that the problem was covered under warranty.
For someone who hadn’t done online research like I did, it would also cause uncertainty in their mind; the customer service agent couldn’t confirm that the repair would be free of charge until they had sent the device back. This meant they would be charged, at the minimum, for the diagnosis and shipping of the device to the service centre and back, if it turned out not to be covered under warranty.

SAP Leonardo and SAP C/4HANA to the rescue

Enter SAP Leonardo's machine learning capabilities with SAP C/4HANA. Now, imagine the same scenario but with the machine learning capability in SAP Service Cloud, which is part of the SAP C/4HANA suite of products.
When I report the issue, the machine learning algorithm identifies existing tickets where a similar issue was reported and how they were resolved. It predicts that the fault is covered under warranty as it is a manufacturing defect and even proposes the parts that are required for the repair to take place. With this knowledge:
  • The customer service agent can provide a reasonably reliable estimate to customer (zero cost in this example).
  • The customer service agent can see that proposed parts are not in stock in the depot receiving the device for the repair and can trigger a request for advanced shipment to the depot, reducing the time it takes to repair.
Taking this further, in the case of a field service scenario, the field service technician can ensure the proposed parts are in their van stock before making the trip to the customer’s premises. This not only saves an expensive trip to the customer's location only to realise that the part(s) needed are not in stock and the technician must pay another visit, but also makes the customer experience much smoother.
Another problem this solves is to do with compliance and it is one I have come across recently with a client. When they use third-party authorised service centres, internal governance policies mandate that third-party vendors should only be instructed to work on their behalf when there is knowledge of a rough indication of the cost, albeit with a high degree of precision.

Currently, as this is something that can only happen once a technician has diagnosed the device - it is extremely difficult if not almost impossible to predict. With machine learning capabilities, however, this will change and the customer service team will be able to predict with reliable accuracy the cost of a repair.

Future developments of SAP Service Cloud

This use case is not yet built but it is on the roadmap. Real progress has already been made in this area and at present, you can use machine learning to automatically propose a category for the ticket. This categorisation can then be utilised to automatically route the ticket to the appropriate team.

In addition to this use case, further capabilities can be imagined, such as automated responses and predictive maintenance of assets, just to name a couple. A lot more can be achieved as machine learning capabilities mature, leading to increased efficiencies and customer satisfaction.

Want to find out more about SAP Service Cloud and machine learning? Watch our latest feature with our Head of Customer Engagement and Commerce Thierry Crifasi.

View comments


Blog post currently doesn't have any comments.

Security code

About the author

Uphaar Agrawalla

Solution Architect - SAP C/4HANA

Bluefin and SAP S/4HANA - welcome to the one horse race

We use cookies to provide you with the best browsing experience. By continuing to use this site you agree to our use of cookies.