Customer Story:
Rhaetian Railway (RhB)

Key facts
Working together
since
2023
Chatbot sessions
12000+
User requests
26000+
Languages
3

Customer Story:
Rhaetian Railway

AI at the Rhaetian Railway

The Rhaetian Railway (RhB) is one of Switzerland’s best-known railway companies, world-famous for spectacular routes such as the Bernina and Glacier Express. With its long and rich history, the RhB attracts travelers from all over the world, who often plan their journeys months in advance. To remain successful in the future, the RhB pursues the clear goal of being a technological pioneer in the railway sector and of using innovation and digitalization to find answers to the most pressing challenges of our time.

As part of this effort, the RhB worked with ParetoLabs, a Microsoft partner startup, to develop Flurina, an AI-powered chatbot based on Microsoft Azure OpenAI Services. Since its launch in September 2023, Flurina has complemented the RhB’s existing service channels and significantly reduced the workload of human staff.

Impressive results after just six months:

  • More than 12,000 chatbot sessions and over 26,000 answered customer questions
  • Available around the clock, in German, English, and Italian
  • Fast and reliable information on reservations, itineraries, and other services
  • Relief for the Railservice team, which can focus more on phone and email inquiries as well as internal projects

Why the RhB relies on Azure OpenAI

Faced with a steady increase in customer inquiries and a shortage of skilled workers, the RhB launched an internal innovation program and evaluated the use of generative AI. A six-week feasibility study with ParetoLabs quickly showed that standard chatbots lacked the up-to-date, company-specific knowledge the RhB needed. With Azure OpenAI Services and Swiss data residency, a tailor-made, secure, and privacy-compliant solution was created.

How Flurina was developed:

  • Identification of the most important customer concerns
  • Development of a comprehensive knowledge base
  • Optimization of search paths
  • Introduction of a fact check to verify links, phone numbers, and email addresses
  • Integration of customer feedback during the beta phase
“As a traditional company, we need technological progress to remain successful in the future and to continue meeting our customers’ high expectations.”
Patrick Grünenfelder
Head of Sales Systems, RhB

Impact on customers and employees

Flurina was enthusiastically received from the very beginning – especially by international travelers and younger customers who prefer self-service. Without major marketing efforts, the chatbot quickly established itself, showing that acceptance of digital services is high today.

Internally, Flurina has also become an important part of the RhB’s modern workplace strategy. It increases efficiency and helps attract qualified specialists.

Next steps

The RhB and ParetoLabs plan to further expand Flurina’s capabilities:

  • Support in processing emails
  • Expansion of the knowledge base
  • Ensuring consistently excellent customer service across all service areas

Customer Story:
Migros Bank

Key facts
Working together
since
2024
Channels
3
User requests per month
15000+
Languages
1

Customer Story:
Migros Bank

AI at Migros Bank – Added value for customers and employees

Migros Bank uses artificial intelligence in a targeted way to make information more easily accessible to customers and employees. Based on the AI platform from ParetoLabs, the digital assistant “Migros Bank GPT” was developed and is already in use in selected areas of the website. Customers can interact with the assistant naturally and receive targeted help, independent of opening hours.

In addition, a Secure AI Chat and an email assistant are in use for employees. Thanks to anonymization and the use of LLM models in the Swiss cloud, the Secure AI Chat offers a secure alternative to public services such as ChatGPT. The email assistant supports client advisors with suitable response suggestions and follows the “human-in-the-loop” principle, so that editorial responsibility always remains with the employee.

Key building blocks of the AI deployment:

  • Development of a central knowledge base (“golden source”) with interest-rate and pricing information, product details, instructions, and information on branches
  • Continuous improvement process: insufficient answers lead to targeted expansion of the knowledge base
  • Definition of clear criteria for when an answer is considered good and when to refer to a specialist
  • Definition of boundaries: the digital assistant exclusively answers questions about the bank and its services and does not provide advisory recommendations

Efficiency and quality at the customer interface
The digital assistant currently processes more than 15,000 customer inquiries per month. It analyzes the inquiries, searches the knowledge base, generates answers, and performs a fact check. Customer inquiries are anonymized before they reach the Swiss cloud; customer-specific information is only added back before the answer is delivered. In this way, Migros Bank ensures that answers are provided within two to three seconds, accurately and in compliance with data protection requirements.

“With ParetoLabs, we achieved tangible results in a very short time. The scalable AI platform delivers measurable added value and impresses with an interdisciplinary team of experts that supports us end-to-end. Thanks to its technological edge as a Microsoft startup partner, ParetoLabs makes an important contribution to our sustainable success.”
Cristophe Makni
Head Data & Automation, Migros Bank

Email assistant for employees
The email assistant is integrated directly into MS Outlook and creates drafts for customer responses at the push of a button. Employees can add case-specific details as keywords and receive precise response suggestions based on them, which they can adjust manually if needed.

Conclusion and outlook
Migros Bank is among the pioneers of productive AI use among Swiss financial institutions. It receives positive feedback, answers numerous customer inquiries efficiently, and is already planning to expand towards “agentic AI” in order to trigger workflows in the future as well.

Contact us.
We support you in every phase of your AI project.

Get in touch and find out what’s possible. We look forward to bringing your ideas to life.

Thank you!
We will get back to you shortly.
Oops! Something went wrong while submitting the form.