Fewer tickets, minimized editorial, maintenance, and service efforts, and more satisfied customers: the benefits of targeted use of ChatGPT in customer service are significant. Increasingly, service organizations are exploring, together with USU experts, the specific use cases that an integrated ChatGPT bot within USU's Multibot solution can cover. The concept of interconnected chatbots – a "USU Bot Universe" – enables the construction of highly efficient dialogue systems with minimal effort. Thanks to the Multibot architecture, generative AI language models like ChatGPT can also be easily integrated. As a result of numerous customer and research projects, USU has many years of experience with Large Language Models (LLM).
For example, an IT service provider uses a ChatGPT bot in its customer portal as a central service module for general IT questions. In addition to traditional chatbots for areas like applications, remote access, and smartphones, the new ChatGPT bot acts as an expert for general IT knowledge. For example, it correctly answers questions about how to connect network drives, how to develop or use macros in Excel, or which keyboard combination creates emojis. In the context of the question, the ChatGPT bot provides the complete IT knowledge of the internet. Its language comprehension allows for refining questions and answers so that ChatGPT can provide concrete help in practice. This benefits numerous employees initially, and in the second step tens of thousands of end users on the customer side.
An insurance company uses the ChatGPT bot in the context of the USU knowledge database for the automated creation of FAQ documents as well. Previously, experts had to read complex documents and formulate questions and answers as clearly as possible – now their effort is minimized. Extensive texts such as contracts and manuals can be used as a basis for clearly written FAQs and their answers.
"We expect that in the next 3-4 years, no customer service software will be able to do without ChatGPT or a comparable AI language model. Accordingly, customer service and its services will completely change with the help of generative AI. Where previously documents had to be created for different purposes, today there are only combinable text modules that need to be made available to the AI language machine. From this, the system generates various documents for different needs. Already now, ChatGPT has also changed the role of the knowledge database, because future texts will have to be written in such a way that they are understood by machines, not just by people," says Harald Huber, CTO and Managing Director of USU.