Case Study: From 16 hours to 30 Minutes – Trask AI Agent Helps Speed Up MONETA Bank's Payment Card Disputes
Modern banks face increasing pressure to resolve payment card disputes quickly, accurately, and in full compliance with Payment Services Directive 2 (PSD2). MONETA Bank, a leading CEE bank (formerly GE Money Bank), can now automate most of this process with an agentic AI solution co-developed with Trask – reducing handling time from ~16 hours to ~30 minutes.
[.infobox][.infobox-heading]Business Impact [.infobox-heading] • Up to 30× faster processing in most cases.
• 97% of cases evaluated successfully.
• Claims handling moved to a 24/7 mode.
• ~5,000 cases resolved per month, with capacity to scale higher. [.infobox]
The situation: The transformation of the claims resolution process increased requests by 70%
PSD2 tightened the rules and deadlines for handling client complaints. At the same time, MONETA Bank opened new digital channels that allowed customers to submit card complaints anytime, from anywhere. This easier intake immediately increased the number of submitted requests by 70%, putting additional pressure on processing capacity and extending response times.
The solution: An AI agent that covers the process – from client complaint to possible disbursement
To meet the urgent need, Trask delivered a proof of concept of the Trask Payments Dispute Agent within one month and took it to production in two. Powered by an Claude Large Language Models served on Amazon Bedrock, the agent understands and evaluates customer disputes and their attachments, selects from 20 possible outcomes, and when appropriate, executes reimbursement and customer communication.
The bank now handles disputes 24/7 at a fraction of the previous cost and time, while ensuring the correct payout and keeping every decision fully compliant, documented, and auditable. This reduces back-office workload and prevents overpayments or incorrect reimbursements.
The process is now completed in ~30 minutes, down from ~16 hours.
An Agent claim process handling workflow
- Dispute submission: A dispute can be raised via Branch, Internet Banking, Call Center, or via Mobile. The client describes the dispute and provides accompanying attachments as evidence.
- Case intake: All details are stored in a unified queue – previously handled manually by the back-office team. Every 10 minutes, Amazon EventBridge Scheduler triggers the agentic process which evaluates any new cases in the queue.
- Advanced reasoning and evaluation: The agent enriches each case, evaluates the claim through advanced reasoning capabilities by state-of-the-art Claude Sonnet models served via Amazon Bedrock, and determines the appropriate outcome.
- Execution & communication: During the process and at the end, the agent calls ~40 existing APIs to disburse the correct amount, notify the customer, log the decision, and save notes for the card team in the unified queue.
The dispute resolution agent has demonstrated how AI can be successfully deployed in a production banking environment and integrated into our existing AWS ecosystem. More importantly, it has created a repeatable blueprint that we can reuse for other operational processes across the bank.
— Denisa Wimmerová Šnebergrová, Senior Manager, Payments & Customer Service, MONETA Money Bank
Want to speed up and automate processes?
Let’s explore how a modular digital ecosystem can work for you. Get in touch.
About MONETA Bank
MONETA Bank, a.s. is a leading bank in the Czech market, offering a wide range of competitive products from deposits to investments and insurance while specializing in credit products for Czech households and loans for entrepreneurs and small and medium-sized enterprises. Formerly known as GE Money Bank, it offers services such as mortgage refinancing, loan consolidation, and online banking.
Read more insights
