Wienka – Customer Support Chatbot for Wiener osiguranje

Overview

Our friends at Wiener Osiguranje Hrvatska have a lot on their plate with designing, promoting, and selling insurance products. But their work only begins there. They also need to keep hundreds of thousands of customers happy, and that heavily depends on how good their customer service is. High-quality, always-available support has become one of the key differentiators in the insurance industry, and that is where Wienka comes in.Over two days, colleagues from across telecommunications, network planning, and ICT services gathered to deepen their understanding of machine learning systems and how to identify where these technologies can make a difference in our business context. We left with a stronger technical intuition, hands-on experience in framing machine learning problems, and a clearer view of how data-driven solutions can support smarter, more efficient operations.

"Our collaboration with CROZ is a great example of one of those rare moments in business when a strong idea meets the right technology. We defined Wienka’s character with a recognizable tone of voice, empathy, smart humor, and a reliable, calm communication style, and CROZ turned that concept into an advanced AI assistant that behaves exactly like that in practice. The result? The first AI assistant in the industry that does not sound like a robot, truly understands its users, and communicates in an authentic, clear, and relatable way."

— Elizabeta Sarjanovic, Product Manager at Wiener Osiguranje

Challenge

It is difficult and time-consuming to maintain a high level of quality in customer support services. Contact centers face limited availability of agents, agent churn, and a slow onboarding process for new employees.

There are also challenges with long response times, accuracy of information, and the availability of up-to-date answers. All of this can lead to missed opportunities, lower customer satisfaction, and frustration on both sides.

Wiener Osiguranje wanted to:

Improve availability of support without endlessly growing the team
Reduce waiting times and increase answer accuracy
Provide a consistent, friendly communication style across all interactions
Keep human agents focused on complex, high-value cases instead of repetitive queries

Solution

The CROZ team offered a modern solution for the constant challenge of not having enough workforce. Wienka, a Customer Support Chatbot powered by AI and built on the IBM watsonx platform, augments human workers by taking over a large portion of incoming queries and completely solving many of the customers inquiries.

Wienka is a well-informed Customer Support Chatbot with access to a large, curated knowledge base. It contains information about various insurance products, internal Wiener practices, and answers to frequently asked questions. When Wienka can answer the question with confidence, she does it in a friendly, carefully fine-tuned tone of voice, so the customer enjoys the conversation instead of feeling annoyed by a generic bot.

When Wienka thinks she cannot give a good answer, or recognizes that the user is getting frustrated, she quickly assigns a human worker to the issue. In this way, the Customer Support Chatbot always keeps customer satisfaction as the top priority and makes the handover to a human agent smooth and natural.

Wienka was developed using IBM Cloud Platform, but a similar Customer Support Chatbot solution could be built using any other major cloud service such as Azure or GCP, or even as a fully on-premises implementation. The service that sets IBM apart is watsonx Assistant, which enables quick chatbot development and simple integration into existing systems, making it easier to scale and improve customer support over time.

Financial impact

Contact center workload reduced, resulting in measurable savings through fewer repetitive inquiries and shorter handling times.
Productivity of sales agents improved, as Wienka enables instant access to verified information and faster contract generation.
The modular architecture significantly lowers future implementation costs, as additional AI assistants can now be deployed using 80% of existing components, cutting future development time and costs by more than half.

Operational improvements

Average response time for client and agent inquiries reduced from minutes to seconds, representing a significant improvement in service speed.
Accuracy of information delivery improved during acceptance testing (772 complex queries), ensuring reliable and compliant communication.
The AI assistant now operates 24/7, improving availability and customer support beyond standard working hours.
Agents can now access insurance policy details, coverage information, and contract clauses in real time, eliminating dependency on back-office support and manual document searches.

Strategic outcomes

Improved digital competitiveness and brand positioning, with Wienka recognized as one of the first generative AI assistants in Croatia’s financial industry.
Significant improvement in employee and customer satisfaction, driven by faster service, engaging tone of voice, and reduced waiting times.
Accelerated AI adoption culture within the client organization, with Wienka serving as a catalyst for future digital transformation projects.