In 2026 process optimisation and automation remain central to banking, but the role of artificial intelligence (AI) is nuanced: AI is not a universally deterministic or fully predictable tool, so banks need to assign clearly defined roles for it in development and automation projects. Human oversight, validation and the possibility to correct AI outputs remain essential in banking operations.
50–40–10: a simple model for practice
A simplified 50–40–10 model illustrates how banks distribute effort in process optimisation:
- About 50%: simplification and rationalisation of processes.
- About 40%: robotisation (RPA) of rule‑based, repetitive tasks.
- About 10%: AI‑based solutions for the most complex challenges where systems must interpret information, recognise patterns and produce decision proposals.
The model highlights that AI does not replace robotisation but complements it where rule‑based automation reaches its limits.
RPA in practice: K&H’s experience
According to the source material from Magnific, K&H’s RPA (Robotic Process Automation) solutions deploy software robots to automate repetitive, rule‑based and high‑volume administrative or back‑office tasks. Typical use cases include data processing in banking systems, report generation, account handling and data checks related to loan assessments. Automation shortens turnaround times, reduces errors and relieves staff of significant manual workload.
Ozorai Dénes, Head of IT at K&H, said that nearly 100 robots work on development efforts at the bank; dedicated process experts collaborate closely with the IT team to lead optimisation projects. Back office areas are at the forefront of robotisation, and customers experience the benefits through faster processing times and quicker service.
Concrete solutions
K&H operates several automation solutions:
- Agent Portal: a robot that supports creditworthiness checks in the background, relieving case handlers and accelerating the process.
- Automation of retail savings account openings: the system verifies eligibility criteria and performs the related administration, making the process faster, more consistent and more efficient.
- Corporate deposit processing: a robot supports handling large corporate deposits, in some cases worth several hundred million forints, ensuring accurate and timely processing while reducing manual workload and operational risk.
These solutions generally follow pre‑defined, controlled workflows and do not make autonomous decisions; they reliably execute programmed steps without human intervention.
Where AI adds value
AI provides added value where recognising relationships, pattern detection or complex decision support is required. Fraud prevention is a typical AI use case: within information security and data protection rules, AI can detect suspicious money flows and identify patterns indicative of abuse, and suggest actions.
For example, if a customer previously provided relatives’ contact details, an AI system could propose notifying relatives in a suspected abuse case — subject to information security and privacy constraints.
Relationship between RPA and AI
In practice, robotisation and AI are complementary rather than competitive. Robots improve speed and reliability for repetitive, rule‑based tasks, while AI supports complex, data‑driven decision situations. Combined, the two approaches increase efficiency, reduce operational risks and relieve staff.
Conclusion
In 2026 banks continue to prioritise process rationalisation and robotisation, with AI applied selectively to the most complex problems. The 50–40–10 model and K&H’s experience indicate that successful automation requires a staged approach: first simplify and rationalise processes, then robotise rule‑based tasks, and finally introduce AI where it provides genuine added value.
Source: Magnific
