Belgian rail operator SNCB has launched Mobi 2.0, a new version of its virtual passenger assistant that uses generative artificial intelligence technology. According to the company, it is one of the first applications of its kind to be implemented on a large scale in Belgium in the public services sector.
The assistant is available 24 hours a day, seven days a week, and provides information to travelers via the SNCB website, as well as on Messenger and WhatsApp channels. Later this year, Mobi 2.0 will also be integrated into the operator’s mobile app.
Significantly improved understanding rate
SNCB has been using a virtual assistant since 2020, which currently handles around 30,000 conversations per month. According to data provided by the company, the previous version correctly recognized around 65% of questions, while Mobi 2.0 is said to exceed a 99% understanding level, even in the case of spelling mistakes or complex formulations.
The new assistant can provide answers in four languages—French, Dutch, German, and English—and is designed to offer more natural interactions that are less dependent on predefined scenarios.
Support for customer service
Mobi 2.0 is primarily intended to handle frequently asked and urgent questions, such as information about routes, connections, delays, or fares. According to SNCB, this approach allows customer service teams to focus on more complex requests that require human intervention, such as special situations or refund requests.
In cases where a request cannot be resolved automatically, the assistant transfers the conversation to an SNCB agent, thus ensuring continuity of personalized support.
Security and use in a controlled environment
SNCB specifies that Mobi 2.0 operates in a closed digital environment, fed exclusively with the company’s commercial information, and includes several levels of protection to prevent the use of inappropriate language.
A pilot project for future AI applications
Mobi 2.0 is SNCB’s first large-scale generative artificial intelligence project dedicated to customer service.
The operator considers this implementation to be a model for future AI applications, with the technology to be expanded further in terms of both functionality and communication channels.
Share on:
