AI embedded into Banking operations

As experts at our banking client, we participate in several independent or partially related AI projects aimed at automating banking processes, supporting and accelerating administrative work, primarily in the areas of customer communication and internal processing. Our task is the precise understanding of business needs and structuring them into technically feasible solutions suitable for development, including the elaboration of alternative solution directions and the preparation of developments.

In complaint handling processes and in the case of inquiries arriving via the Contact Center e-mail channel – typically aimed at general information requests – AI supports the categorization of messages, faster task assignment to the competent administrators or resolution groups, and the creation of draft response letters during complaint handling with human approval. In addition, we support the introduction of the BIAN framework with AI-based document processing: we identify functions from banking and IT documents and classify them into service domains, assisting the analytical and planning work of architects.

We are also participating in a research project jointly with the staff of ELTE, the aim of which is to create software that measures the performance of AI agents, uncovers their errors, and is intended to provide a cost and efficiency comparison of different solutions.

The implemented solutions have delivered measurable business results. In complaint handling processes, manual workload was reduced by 5.5 FTEs, resulting in savings exceeding HUF 300 million, while complaint resolution times were significantly shortened. In the Contact Center e-mail processing area, a reduction of 10 FTEs was achieved, with an identified business impact exceeding HUF 300 million over a five-year period.