Aristech provides multilingual, data-protection compliant voicebot technology
Air France-KLM automates customer service with a European AI solution
The airline AirFrance-KLM is relying on a European solution to expand its customer service. Heidelberg-based language technology provider Aristech is supporting the group with a multilingual AI-powered voicebot for customer service, which is gradually being rolled out in European markets since July 2025.
The solution is currently available in Germany, the United Kingdom, Belgium, Luxembourg, Austria, and Switzerland. At this time, the voicebot accepts calls in German, English, French, and Dutch. Additional languages are in preparation. Worldwide deployment is planned for the future.
The AI-supported voicebot replaces conventional call center menu navigation with dialogue-based speech recognition. It identifies the callers’ concerns, records relevant information such as booking or ticket numbers and details of baggage issues, and passes this on in a structured manner to the responsible service employees. This greatly reduces processing time, as the right person is contacted directly and service employees can start conversations with the specific concern and the recorded caller data already at hand.
“With the Air France-KLM project, we are demonstrating that AI-supported customer service works reliably, scalably, and in compliance with data protection regulations, even in highly complex, international environments,” says Carolin Edler-Mende, CEO and founder of Aristech.
Data protection and data sovereignty as key criteria
The decisive factor in the awarding of the contract was the search for a sovereign European solution that meets high standards of performance, data protection, operation, and hosting. Air France-KLM retains sovereignty over its own data at all times and can specify exactly what information is collected, where it is stored, and how long it is retained.
The solution is designed for use in situations with highly fluctuating call volumes. In the aviation industry, high call volumes occur regularly, for example in the event of weather-related disruptions, strikes, or last-minute rebookings. The voicebot developed by Aristech can operate reliably even with over 1,000 callers at the same time.
Language diversity as a challenge
A particular challenge is the linguistic diversity encountered when interacting with customers in an international environment. The voicebot currently processes different languages and varieties, including Swiss German, Austrian variants of German, and Belgian French. In addition, many callers use national, international, or individually chosen alphabets when spelling (such as “A for auto”), often in a language that is not their native tongue.
Recognizing alphanumeric information such as booking or ticket numbers is particularly challenging in this context. Letters and numbers are pronounced or spelled differently depending on the language. Many users mix letters, words, and phonetic spellings within a single dictation. In order to process these real-life speech patterns correctly, Aristech optimizes the speech recognition models used specifically for concrete tasks and scenarios. Aristech is also supported by the French company Illuin, which is acting as a subcontractor in this project and whose technology is being used in the field of natural language processing (NLP).
How AI can further change customer service
In the future, airlines will be able to use AI-powered voicebots to automate service processes either partially or completely. This includes, for example, rebookings, changes to existing bookings, or the processing of combined requests such as flight route, seat, or meal changes. This requires deep integration into the airline’s existing core systems, particularly reservation, ticketing, and billing systems.
Through the use of generative AI, these systems can understand customer requests in context. This allows standardized, but previously labor-intensive service cases to be handled automatically, while human service teams can focus on special cases, escalations, and exceptional situations. The primary benefit comes from the close integration of language interfaces, generative AI, and clearly defined, automatable business processes.
picture: freepik
