Full steam ahead to a digital future: The Flurina chatbot based on Azure OpenAI Services

Using artificial intelligence to chart new paths in customer care

Recent years have seen a constant rise in the number of customer inquiries coming into RhB’s Rail Service unit by phone and in writing. The extra work for staff was then made worse by the general shortage of skilled labor. RhB needed to act: In the context of an internal innovation program, it quickly came up with the idea of using generative artificial intelligence (AI) to develop a groundbreaking solution in the shape of an intelligent chatbot for Rail Service activities. The company took this idea to Swiss startup ParetoLabs, which immediately set to work on a six-week proof of concept phase to test whether existing language models were suitable for use in customer care. It quickly became apparent that standard chatbots do not meet RhB’s requirements, as they possess neither up-to-date nor specific information about the company. For this reason, RhB and ParetoLabs decided to use Azure OpenAI Services to develop a chatbot tailored specifically to the company’s needs. 

Made to measure: The Flurina chatbot, based on Azure OpenAI Services

ParetoLabs developed an initial version of the digital assistant Flurina in several phases. To begin with, it was necessary to group together customers’ most important concerns, produce a knowledge database as the platform for the AI chatbot and optimize information search pathways. Within a few weeks, a beta version of the chatbot was rolled out ready for testing by RhB’s customers. During this test phase, feedback from users allowed errors and possible improvements to be identified. For example, it was discovered that, in spite of all the precautionary measures taken, the chatbot at times created links that were non-existent. In response, ParetoLabs developed a fact checker that reviewed and, if necessary, corrected links, telephone numbers and e-mail addresses before passing them on to the customer. Once this kind of problem had been remedied, Flurina was able to go into service in customer communication. Since then, RhB has been reaping the benefits of a productive chatbot that runs successfully on the Azure OpenAI platform at a Swiss data center in accordance with the strictest security and privacy stipulations.

RhB customers like Flurina

Communicating with customers in German, English and Italian, the chatbot was well received from the outset. Available to everyone around the clock, it is quick and easy to use, providing reliable information about reservations, travel itineraries and many more aspects besides. International customers in particular enjoy interacting with Flurina. Experience shows that many potential customers shy away from placing phone calls with foreign numbers, preferring instead to submit inquiries online. The company’s younger customers typically favor self-service options; and they too have responded very well to chatbot Flurina. Across all age groups, it has become apparent that the willingness to use digital service offerings is now high. In the shape of virtual assistant Flurina, RhB is scratching exactly where people today itch.

Toward the ‘modern workplace’ – with digitalization and artificial intelligence

The staff of RhB are likewise benefiting from their new chatbot colleague, available to them via Azure OpenAI Services. The digital service channel eases the workload above all for the Rail Service team, giving them more time to devote to inquiries received by phone or e-mail, as well as to other projects. At the same time, the chatbot also supplies employees in other parts of the company with all kinds of relevant information.

The digital assistant thus plays its part in crafting the ‘modern workplace’ that RhB is advancing as part of its digitalization strategy. Greater efficiency and agile work processes are two good ways to attract suitably skilled people on the labor market.

RhB to further expand its digital customer care

Going forward, RhB plans to further optimize the AI chatbot Flurina, again in collaboration with ParetoLabs, the youthful, ‘Microsoft for Startups’-supported company. The next step is to conduct a detailed analysis in order to exactly quantify the benefits and efficiency of the digital assistant. Moving on from there, the chatbot will be fed – via Azure OpenAI Services – with even larger volumes of data. The aim? To guarantee the same high quality of service across all aspects of the company’s activities.