ChatGPT, or "Generative Pre-trained Transformer", is a deep-learning model trained on a large amount of text data to produce text that is similar to text spoken by humans. Essentially, it is a computer program that is able to understand and respond to natural language input to mimic a human conversation. Another important capability of ChatGPT is to understand the context of a conversation. It is able to take into account what has been said before and respond in a way that makes sense in the context of the conversation. This is a crucial step in enabling natural, human-like interactions with computers. And thus ChatGPT or other Large Language Models (LLMs) based on the same principle are indeed an excellent working tool in customer service.
Instead of a shortage of skilled workers: intent recognition and automated processing
Chatbots and voicebots based on this technology can answer general customer questions, solve problems and even handle more complex requests in a natural dialogue. The customer does not feel like they are talking to a robot. This can help build loyalty and trust in a brand. Such bots give quick and accurate answers around the clock. This shortens waiting times for customers and increases customer satisfaction. In view of the severe shortage of specialists in customer service and contact centres, conversational automation technologies are a really promising solution for permanently relieving the strained personnel situation.
Caution: Personal data!
However, two aspects must generally be taken into account. If the customer's personal data is processed, the regulations of the GDPR apply. This means that companies must ensure that the LLMs used comply with these regulations.
The second aspect concerns the specific data or information of the company. The chatbot can only give a satisfactory answer about the status of a delivery if it has access to the corresponding systems. Internal company guidelines, customer-specific data from the CRM system, logistics or invoice data must also be integrated for conversational automation if more than just a superficial dialogue is to be achieved. If a customer requests a credit note in dialogue with the chatbot, for example, the credit note and the transfer itself can be triggered automatically. So it quickly becomes complex if a case is really to be processed – and it should be!
Real goldmine: Automated intent analysis
However, conversational automation not only promises great efficiency gains in dialogue with customers. So far hardly used and perhaps underestimated is the potential of intent analysis. Whether ChatGPT, another LLM or a semantic network is used is a company-specific decision. The result is a goldmine: an intent analysis provides information about the main drivers of call and contact reasons. And this is where there have been huge problems in practice so far. The recording in the contact centre is error-prone, not representative and not very meaningful, especially if the main call reason is "other". However, if you know the contact drivers, you can influence them and improve online offers, adapt training and coaching programmes, update knowledge management and optimise product management.
When the employee wins, the customer wins: copilots support
Employee Experience means: What helps the employees helps the customers. How can conversational automation support this? It is ideal if the service staff can concentrate on the dialogue – telephone calls, chats – and busy digital helpers work in the background. Knowledge management is an ideal field of application for Conversational Automation. The automatic retrieval of suitable answers hidden in PDFs, databases, Word and Excel files, in the intranet, saves time in the conversation, as does automatic information on complex questions.
It is also helpful to show the employee the context from which customers are acting and what meaningful answers or offers could look like. The service worker then still decides for him/herself what is to be used in the dialogue and how. Such copilot solutions are not about disempowering employees or even replacing them, but about support.
Conclusion: So far, Conversational Automation has only scratched the surface of the possibilities. But this is already so impressive and promising that one can look forward to further applications!
