In customer service, various methods and tools have long been used to supply the basic and up-to-date knowledge necessary for the work of customer service employees. The spectrum of methods ranges from quite simple solutions to applications developed specifically for this purpose.
With the advent of omnichannel solutions, it quickly became clear that knowledge must also be made available to customers in the form of online service FAQs or online chat bots. It is important that there are no discrepancies in the statements between the customer service representative and the information supplied online.
For this purpose, knowledge management editorial offices were set up to obtain procedural and up-to-date knowledge and make it available in the systems.
The cleansing of the systems from old, outdated information is often inadequate. Often, the presentation of the information is also in need of improvement – in the case to open the content, it is often necessary to click and scroll several times. Just as often, „if-then“ conditions and test procedures in different systems are described for one process.
Overall, this means that customer service employees at best read the information marked as „new“. During the induction phase, new employees are often overwhelmed by using the agent frontend and the knowledge management interface at the same time when talking to customers.
How useful would it be if the employee did not even have to open the knowledge database to research all the relevant information about the process or even get a hint in case of an incorrect entry in a field of the agent frontend?
To achieve this, a robotics solution can be used. Since it is about support in an ongoing customer conversation, a real-time robotics solution becomes necessary. The robotics solutions in this regard are referred to as „Real-Time-Interaction-Management (RTIM Bot)“, „Attended RPA-Bots“ or „Robotic Desktop Automation (RDA-Bot)“.
With these solutions, the focus is on bot-human interaction. The RDA-Bot takes on the role of a digital assistant that supports the employee almost in real time with recommendations for action or relevant information from different systems.
Unlike RPA, RDA works real-time on the employee’s desktop. Regarding the best use of the knowledge management system, there can be four main solutions:
For solutions one and three, the use of AI, for example a ChatGPT solution, can be of great benefit:
Up-to-date information can be summarized by ChatGPT. Here, ChatGPT can also summarize the old information from the knowledge management system and compare it with the daily updated information.
An open-ended question can also be answered by ChatGPT with compressed information that is much shorter than the original entry in the knowledge management system. ChatGPT is much more efficient here than the search function in the knowledge management system.
Of course, ChatGPT must first learn the contents of the knowledge database to use it.
In addition to the process help described above, the RDA-Bot can of course also be used for any form of „next-best-activity“ help or for sales purposes as „cross- and upsell“ support.
Overall, the use of an RDA bot leads to a significant reduction in processing errors and a shortening of processing time. Likewise, the learning curve for new employees can be significantly reduced and the success in cross-selling and upselling can also be increased.
Of course, the described RDA functions can also be used in the online service to support the processes to increase the usage rate and avoid calls to customer service.
Open customer questions in the online service can be resolved by a chat bot. Here, too, ChatGPT can lead to a significant increase in user-friendliness.
Hans-Joachim Grün – Senior Consultant
junokai