How companies can use AI chatbots efficiently and compliantly

AI chatbots for companies have the potential to transform the way we communicate with internal and external stakeholders. Applications range from customer service and knowledge management to supporting maintenance processes. The white paper "Smart Assistant. Company-internal AI chatbots - individual and legally compliant" from Lufthansa Industry Solutions shows practical ways in which intelligent assistants can already be used today.

Norderstedt, October 17, 2024 – The white paper "Smart Assistant" shows how companies can use the potential of intelligent assistants with Generative AI. It offers strategies for increasing the efficiency of business processes while complying with the legal requirements of data protection and the EU AI Act.

Our experience from customer projects shows that companies not only gain competitive advantages, but also sustainably strengthen their innovative power.

Theresa Tran,
Business Manager AI & Data Analytics at LHIND

"In addition, our experience from customer projects shows that companies that rely on AI assistants and in-deep automation not only gain competitive advantages, but also sustainably strengthen their innovative power," says Theresa Tran, Business Manager AI & Data Analytics at Lufthansa Industry Solutions (LHIND).

Data quality determines the success of AI

The quality of the underlying data plays a critical role in the success of AI chatbots. Companies need to ensure that the data for the chatbot is of high quality and legally sound. This includes unstructured data from documents, databases and other digital sources.

With the new Retrieval Augmented Generation (RAG) technology, AI chatbots can access specific data and provide accurate, fact-based answers. This allows organizations to use their own data as the primary source for the chatbot. This method prevents the AI from providing incorrect information. With RAG, companies can tailor their chatbots to meet specific needs. They can then provide not only standardized answers, but also specific, contextual solutions in accounting, logistics, or customer service.

Compliance: AI systems and EU regulation

The LHIND white paper also looks at the legal dimension of AI. Its use is subject to strict regulations, especially since the introduction of the EU AI Act. This law divides AI systems into different risk classes and specifies the conditions under which they may be used. Companies must pay attention to these risk classes when using AI in customer transactions and, under certain circumstances, restrict certain functions.

The company must also implement a functioning data governance.

Max Pillong,
Business Director Artificial Intelligence & Data Analytics at LHIND

"The company must also implement a functioning data governance," says Max Pillong, Business Director Artificial Intelligence & Data Analytics at LHIND. "It must ensure that only suitable or legally unobjectionable data is used, i.e. no personal data that falls under data protection, for example."

The introduction of such technologies requires not only technological adjustments, but also a change in working methods and corporate culture. That is why employee training is important. They need to learn how to communicate effectively with AI chatbots and what questioning techniques to use to get the best results.