At the Banking Technology conference’s first panel, industry leaders argued that in the age of AI the primary constraint for bank innovation has shifted from technology itself to human capability. Panelists said that artificial intelligence paradoxically increases the value of business knowledge and the human dimension in development projects.
Continued investment needs: data, cyber and AI
Balázs Vinnai, president of IVSZ and chief presidential adviser to MBH Bank, said the banking sector currently spends more on windfall taxes than on innovation. He argued we are moving from a development-intensive era to one focused on efficiency gains, while investment must nevertheless remain strong in data assets, cyber security and AI. Vinnai identified data usage, corporate banking and the payments ecosystem as areas with the largest potential for improvement.
Funding, prioritisation and the role of money
András Sebők, Deputy CEO of OTP Bank, noted that having abundant resources does not by itself guarantee good outcomes. OTP has sufficient funds for innovation, but often tighter resource constraints produce better results. Panelists also observed that business leaders now better understand technology, and cross-functional understanding within organisations has improved.
András Kaliszky, incoming Deputy CEO for IT and Operations at Erste Bank Hungary, said that while available funds for development are stagnating, the share of spending directed at innovation is increasing. This makes prioritisation a key management task; most organisations now operate agilely, so leadership faces similar constraints across teams.
AI’s impact on work and development cycles
Panelists gave figures to illustrate AI’s reach: nearly 7 billion people encounter AI daily, a few hundred million use it as a search substitute, and roughly 20 million leverage AI to boost work productivity. Despite this penetration, there is not yet widespread skill in collaborating effectively with AI; business knowledge, programming ability and analytics expertise have grown in importance.
Kaliszky emphasised efforts to use AI to improve life cycles across jobs, with coding only one target area. The main difficulty is enabling the whole organisation to support development-driven progress.
Fast-moving models, control and costs
Sebők warned that development is moving at unprecedented speed and AI models can become outdated in days or weeks — a pace bank compliance functions may struggle to follow. If a use case’s underpinning technology becomes obsolete during deployment, it must be replaced, so organisations need to adopt what he called a shift toward "infinite order" to avoid unmanageable complexity from disorder.
On costs, Vinnai argued that sometimes stepping back is required to move forward, and banks must maximise resources for development. Growing regulation worsens the challenge while market pressure increases. Sebők added that banks tightly control costs but surprises occur, for example rapid rises in token prices.
Kaliszky noted a clear push to increase "change the bank" investments (developmental spending) relative to "run the bank" operational expenditures, and said that spending driven by regulatory compliance is at least no longer increasing.
Key takeaways
- AI can accelerate development and, in successful AI-native transformations, deliver roughly 25–30% end-to-end efficiency improvements, but only if the necessary human skills are available.
- The current bottleneck is human capability: there are relatively few true AI experts in Hungary, and shortages exist in digital marketing, analytics and customer intelligence roles.
- The banking sector must prioritise funding for data governance, cyber security and AI while management focuses on effective resource prioritisation.
The panel made clear that technology alone will not produce breakthroughs — success depends on the combination of human expertise, business understanding and organisational preparedness.
