the challenges of an effective CSM solution

Customer Service Management (CSM) is an extension of IT Service Management (ITSM) for external clients. Whereas the provision of help desks to customers through the use of call centres has been on-going for at least 30 years, the implementation of full-fledged service desks has not followed suit. For the last 20 years, most large organisations have deployed service desks that provide numerous self-help mechanisms for their internal clients. These mechanisms include multiple interaction points such as phone and emails, and also FAQs, alerts, service level information and expert advice. These are all accessed through a portal on the organization’s intranet. CSM is about offering the same level of service to external clients. CSM must not be confused with CRM (Customer Relationships Management) despite the acronyms being very similar. A CRM is used only by internal clients to improve the relationships with external clients. The external clients are not involved. If, however, the CRM is extended to a portal, it becomes a CSM.

The risks that accompany CSM are commensurate with the ambitions. They are mainly the result of service quality requirements that are much more stringent for external clients.

Consider availability and trustworthiness. Internal clients may be a bit frustrated if the service desk portal is a bit slow sometimes, if the response time for their requests takes longer than they would expect or if the answers they are given are not always correct. They may complain to their colleagues or their managers, but the issues will most likely be contained within the organisation.

Now consider what it would be like for external clients. If, say, your insurance company requires you to go through a customer portal to receive service and that portal is too slow or doesn’t respond when you need it, you are likely to spread your frustration to your families, friends and maybe even regulators. More troubling, if the expert advice provided by the portal is wrong, you may end up in trouble even leading to legal procedures. This advice may be wrong for many reasons, mainly because of low data quality, in the organisation’s databases, from external data sources, and the wrong interpretation of this data.

The cost of the portal running 24/7 around the sun with expert advice is often prohibitive, which is why automation comes into play. It enables organisations to substantially lower the cost of the portal by running chatbots powered by generative AI.

The quality of the information provided to external clients is therefore based, ad-minima, on:

• The quality of the data the organisation holds about the services it provides
• The quality of the data the organisation holds about its clients
•  The quality of the data pulled from external sources
•  The quality of the generative AI tools used to identify and summarise this data
• The quality of the data used to train these tools
• The human experts who oversee the process and results

With the high level of the requisite quality, the risks to bad reputation or litigation are not to be dismissed. A fairly recent example (Moffatt v. Air Canada), where a chatbot gave false advice to a customer, who eventually sued the company is a case in point.

Similarly, in Europe one must abide by the EU AI Act and the Cloud Act. This, however, should not be a call to inaction because of the risk of lagging behind technological progress and competition.

In one of our recent projects we helped a public administration to develop and deploy a citizen portal where we had to address all  these aspects. Our approach is described in the framed text.