bunq
genAI financial assistant (search + support)
Finn, the GenAI assistant of the Dutch bank bunq launched in December 2023, handles about 97% of support activity with 90% satisfaction and a 47-second response time, on a base of more than 20 million European users at the end of 2025.
Key points
- Finn GenAI financial assistant that replaces the app's search and absorbs support.
- Third-party LLM connected to transactions via an in-house RAG layer, in the bunq app.
- About 97% of support handled, 90% satisfaction, 47-second response.
- Evidence level B, confirmed active status.
Objective
Replace the app's search function with a conversational assistant that answers directly on finances, budget, and transactions, and absorb support to serve a fast-growing European base without growing the teams.
The deployment
bunq launched Finn in December 2023, presenting itself as Europe's first bank with a GenAI assistant. Finn replaces the app's search function: the user asks questions in natural language about their finances, budget, transactions, and spending. The bank then made Finn fully conversational (memory of previous exchanges, chained questions). According to bunq, Finn independently resolves up to 40% of requests and assists 35% more; on its first anniversary, the bank announced that Finn handles about 97% of support activity with 90% satisfaction and a 47-second response time. bunq had 11 million users in the EU at launch, more than 20 million at the end of 2025.
Results Proof B
Figures published by bunq in its official releases (T1) and picked up by the fintech press (T3/T4), consistent. Operational metrics provided by the brand, not audited as financial results, hence B.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- llm LLM tiers (fournisseur non nomme) bunq does not publicly disclose the provider or the model version behind Finn; the record stays at the documented level: a third-party generative LLM.
- outil Plateforme GenAI maison bunq (RAG) Internal pipelines that connect the LLM to the client's transactions and account data, with memory of previous exchanges.
- plateforme App bunq Finn's single channel, which replaces the in-app search function and absorbs support.
How it runs, concretely
For ops teams-
1Question in the app client
The user asks their question in natural language instead of a search (budget, transaction, spending, support).
-
2Contextualized answer AI
Finn combines the LLM and the client's account data (RAG) to answer precisely, keeping track of previous exchanges.
-
3Support escalation AI / support
Requests that Finn does not resolve are assisted or passed to a human.
-
4Monitoring data team
The team tracks resolution rate, satisfaction, and response time to adjust Finn's scope.
The share of requests resolved without a human and satisfaction. Since Finn relies on the client's transaction data via RAG, answer quality depends directly on the freshness and structure of that data.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- Structured and fresh transaction and account data
- Support knowledge base
- Conversation history for refinement
Org prerequisites
- In-house or tooled RAG layer connecting LLM and client data
- Escalation rule to support
- Owned AI transparency on the product side
Possible stack
- Generative LLM (third-party or open)
- RAG pipeline on the banking data
- Integration with the mobile app
The plan, step by step
- Step 1Scope a set of frequent questions (balance, budget, transactions) and the data access rules: read-only, confidentiality, AI transparency.Deliverable: Functional scope and compliance framework validated.
- Step 2Build the RAG layer between the LLM and the structured account data, and test it on the set of questions in a test environment.Deliverable: Assistant answering the scoped set of questions correctly.
- Step 3Replace the app's search with the assistant for a beta segment, with the AI nature displayed to the client.Deliverable: Beta in production with satisfaction tracking.
- Step 4Open first-level support with a clear escalation rule to a human.Deliverable: Assistant handling simple support requests, working escalation.
- Step 5Measure the without-human resolution rate, satisfaction, and response time, then broaden the covered scope.Deliverable: Resolution dashboard and expansion roadmap.
First step: Replace the app's search with an assistant connected via RAG to the client's transactions, on a scoped set of questions, before opening support.
The same pattern, proven in other industries
ADT
genAI customer agent
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Sources
- S1 bunq becomes the first AI-powered bank in Europe as it unveils its own GenAI platform Primary archive pending
- S2 bunq launches smarter, more powerful upgrade to GenAI financial assistant Primary archive pending
- S3 bunq's GenAI assistant Finn is now fully conversational Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.