bunq Finn Handles 97% of Customer Queries: How an AI Agent Transformed Banking Service

The implementation of the bunq Finn AI agent has enabled bunq, one of Europe's largest financial holdings, to process 97% of customer queries with an average response time of 47 seconds. This significantly surpasses industry benchmarks and demonstrates how AI agents can radically change approaches to customer service in the financial sector.

In the banking sector, routine tasks are often invisible, but they consume a large portion of the budget: thousands of hours are spent on document retrieval, reconciliations, preparing boilerplate responses, and checks that create nothing but merely keep the process afloat. Multiply these hours by a specialist's rate, and it becomes clear how much "just manual work" costs. This has long been solvable; the question is who in your niche will solve it first.

The Genesis of bunq Finn: From Search to Dialogue

Initially launched in December 2023, bunq Finn was a GenAI platform that replaced the standard search function in the bank's app. Its task was simple: to allow users to ask questions about spending, savings, transactions, and app functionality in natural language. Instead of navigating through menus, customers could ask: "What's my average monthly grocery spend?" or "How much did I spend on Amazon this year?" This made financial information more accessible and marked the first step towards deeper AI integration in customer service.

It's important to note that Finn was not created as a narrow ticket-deflection tool. It started with a customer's need to locate, interpret, and act on personal financial information. This foundation laid the groundwork for the broader role of a financial AI assistant that bunq describes today.

Expanding Functionality: From Search to Full-Fledged Support

Since December 2023, bunq Finn has significantly expanded its capabilities, evolving from a simple search tool into a comprehensive service and insight layer. An update in late 2025 showed that Finn could provide more accurate answers, engage in more human-like conversations, support 38 languages, and assist with questions ranging from cards, payments, and savings to merchant information and travel recommendations.

bunq reported that Finn has already answered millions of user queries and achieved a 90% user satisfaction rating. The assistant resolves support queries in an average of 47 seconds, while the industry benchmark is around one minute. Requests requiring deeper knowledge are directed by Finn to the appropriate human specialist, ensuring a seamless transition from automated resolution to expert assistance.

Multilinguality as Part of the Customer Experience

Support for 38 languages makes Finn not just a localization feature, but a key element of the customer experience. The assistant can translate the bunq app and provide real-time speech-to-speech translation during conversations with the support team. For customers who travel, work internationally, or manage money across borders, this significantly simplifies explaining urgent issues in a non-native language. This builds trust and comfort, which is especially important in the sensitive financial sector.

AI Agent Architecture: A Multi-Agent Approach

bunq transitioned to a multi-agent architecture to make its AI-powered customer support system more adaptable to growing service requirements. Initially, a router-based model was used, which became complex to manage as adding new specialized agents created routing complexities, overlapping capabilities, and potential bottlenecks.

The revised design uses an orchestrator that routes requests to a limited number of primary agents. These primary agents can call specialized agents as tools when needed. Instead of requiring a central router to predict every possible customer need, the system allows primary agents to request specific capabilities, such as transaction analysis, knowledge retrieval, or account support, during the interaction. For the customer, this means less visible complexity and a more coherent interface, even if the query requires different types of expertise.

Implementation Results

These impressive figures indicate that bunq Finn not only automates routine tasks but also significantly enhances the quality and speed of service. Automating 97% of inquiries frees up substantial human support resources, allowing specialists to focus on more complex and non-standard tasks requiring empathy and in-depth analysis.

How to Implement This in Your Business: Lessons from bunq

bunq Finn's experience shows that a successful AI agent implementation strategy begins with a clearly defined customer problem, scales through a flexible architecture, and maintains human accountability as capabilities grow. Technology becomes valuable when it makes customers feel more informed and less burdened, not merely when it reduces manual work:

• Start by addressing a specific problem. Finn began by improving search and access to financial information, not by attempting to fully replace human support. This allowed for gradual functional expansion.
• Utilize a flexible multi-agent architecture. This enables the system to adapt to new requirements, add specialized functions without overloading a central router, and ensure seamless customer interaction.
• Maintain the option for human escalation. Finn handles 97% of queries, but it's crucial that customers can easily connect with a live specialist when needed. This isn't just a "fallback" but a trust-building feature.
• Focus on multilingualism. For companies with an international audience or customers who interact in different languages, supporting multiple languages is critical for creating a truly inclusive and effective service.

If this case sounds like what's happening in your company, our manager can help: he'll analyze your business and niche for free and point out where an AI agent would bring a real result in your case. Message the manager