Banks Are Changing the Architecture of Their Business

Banking has undergone a profound change over the past decade, though not so much in its core role as in the way it must operate within the digital economy. Clients still expect the same things from banks as they did before: secure payments, loans, savings, and investments.

However, the market in which these services are offered is now significantly different, shaped by pressure from neobanks, technology companies, increasingly stringent regulation, and rising expectations regarding the speed and simplicity of the user experience. In such circumstances, banks can no longer rely solely on size, trust, and their established position, but must demonstrate operational agility and technological readiness.

In an interview with ICTbusiness Media, Bálint Fischer, Partner and Member of the Management Board at Dorsum, explains why the real transformation of banking is now moving from the front end to the back end, where artificial intelligence is already delivering measurable value, how regulation simultaneously protects and constrains the market, and why the relationship between fintechs and banks is increasingly evolving toward a model of simultaneous cooperation and competition.

How fundamentally has the role of a traditional bank changed over the past decade as digital-first experiences became dominant?

I wouldn’t say the fundamental role of banks has changed a lot lately. Banks are still there to move money, provide loans, safeguard deposits, and help clients manage their investments. What has changed significantly is the environment in which they operate.

Traditional banks are no longer competing only with each other. They are now facing a new generation of challenger players such as Revolut, Robinhood, and other neobanks, that entered the market not as banks, but as technology companies first, with a clear tech-first mindset and strong focus on user experience.

The client needs mostly have not changed. Customers still want to pay, borrow, save, and invest. However, fintech players often deliver more seamless experiences in commoditized areas such as payments and card services. In contrast, more complex, higher value-added services (e.g., mortgages, wealth management or personalized advisory) remain harder to disrupt, and traditional banks still lead in these fields.

Banks rely on experience, capital, and established client bases, but often struggle with legacy systems and less agile operations. Fintechs are technologically flexible but must build scale, regulatory strength, and trust. What we are seeing is a convergence: banks are becoming more like tech companies, while fintechs are becoming more like banks. I am convinced there will be winners and losers on both sides, and they will be the firms that effectively react to new challenges and implement their strategy.

How do you assess regulation as both a safeguard for stability and a potential brake on innovation in banking?

Regulation in banking has a dual role: it is both a stability safeguard and a potential constraint on innovation.

On one hand, regulation protects market integrity, clients, and systemic stability. Deregulation can boost growth up to a point, but beyond that, it increases exposure to misconduct, weak governance, and financial instability. Over our 30 years of operations, we have seen multiple times how lax regulatory environments can enable problematic behavior.

At the same time, excessive regulation reduces competitiveness and slows innovation. When compliance burdens become disproportionate, firms face higher costs and slower time-to-market, which weakens a region’s attractiveness.

A particular challenge in the EU is regulatory arbitrage. Because licenses can often be passported across member states, firms can often choose to operate under more lenient regimes while still accessing stricter markets. This creates an uneven playing field in my view: jurisdictions with tougher standards may lose competitiveness without fully eliminating risk, since less strictly regulated firms can still enter their markets. This is definitely something to work on in the coming years.

AI is already deeply embedded in banking operations — where is it delivering the most tangible value today?

AI in banking can be viewed in two broad categories: traditional machine learning-based (ML) solutions and generative AI-based solutions (GenAI).

Machine learning has been embedded in banking operations for decades. It works best in highly structured environments where large datasets can be analyzed. Typical use cases include fraud detection, anti-money laundering (AML), credit risk modeling, and transaction monitoring. These systems improve speed, accuracy, and security by identifying anomalies or predictive signals in massive volumes of data. Of course, these solutions are constantly evolving but have been with us for a long time.

Generative AI, by contrast, is a recent development that allows us to automate processes that are less structured and heavily human-driven. The most prominent use cases include drafting mortgage contracts, handling iterative client communication, generating personalized investment advice, or constructing model portfolios. By helping to increase the efficiency of fragmented, judgment-based workflows, GenAI enables banks to provide a higher quality service and better user experience to clients, but this process is only getting started, in my opinion.

Looking ahead, we see many banks expect revenue growth without proportional increases in CapEx or OpEx over the next 5 years. This will only happen if GenAI use cases are scaled successfully.

Many banks have achieved parity with neobanks in mobile and front-end experience — why is the real battle now in the backend?

In the last 5-10 years, banks focused on the front end. They developed better apps, offered smoother UX, introduced digital onboarding. That worked for a while, but they have now realized that the real bottleneck is the legacy infrastructure they are sitting on.

Today, the core systems of traditional banks are rigid, not modular, hard to integrate, and have limited scalability and flexibility. This is hardly a surprise, since many of these systems are 15-25-year solutions, but these limitations increasingly curtail banks' ability to move quickly in a rapidly changing environment, even when they use AI. For banks with a rigid architecture, AI can only serve as a useful patch, not a game-changing multiplier.

That is why banks increasingly need modern, cloud-ready, modular, and microservice-based back-office systems to be able to compete with fintechs on every front. Nowadays, satisfying a new client's needs usually takes 3-12 months due to limitations imposed by banks' outdated legacy systems, which is why banks are realizing that improving their front end was just the tip of the iceberg. And one thing about icebergs is that while their tip is the most visible part, most of the ice is underwater. That is where attention needs to shift now, and that is why I believe the real work for banks is only beginning.

What problem are central bank digital currencies really trying to solve from a banking and societal perspective?

CBDCs are often discussed as one concept, but the two main types solve very different problems, so it is worth separating them clearly.

Wholesale CBDCs target interbank money movement and settlement. Today’s cross-bank settlement infrastructure is built on complex legacy architectures. A wholesale CBDC built on blockchain and DLT could modernize that layer by making settlement faster, simpler, and potentially cheaper. In that sense, it is largely a technology and market-infrastructure upgrade that could create meaningful value for all participants.

Retail CBDCs are what most people mean by “CBDC”. They are essentially digital cash issued by the central bank and held by individuals in blockchain-based digital wallets. At this point, this is less a technology question and more a monetary-policy and macroeconomic topic.

That is because retail CBDCs would represent a step towards a one-tier banking system, which raises major questions about deposits, stability, and the role of the two-tier system used today. If retail CBDCs are pursued, technology would then become an important consideration, but today, economists are best equipped to comment on their purpose and decide if they should be introduced.

How has the relationship between fintech companies and banks evolved from disruption to collaboration?

Fintech is really an umbrella term covering very different business models. While public perception focuses on large B2C players like Revolut, Nubank, or Robinhood, they represent only a small share of the ecosystem.

Roughly 10 percent of fintechs operate primarily in a direct-to-consumer (B2C) model. The remaining 90 percent are B2B providers, supplying banks with technology, infrastructure, or specialized services. For that larger group, banks have always been natural partners rather than competitors. In that sense, collaboration is not new.

What has evolved is the scale and impact of the leading B2C fintechs. The strongest among them (the ones mentioned above) have grown significantly and now compete more directly with traditional banks for customers, deposits, and engagement, making the competitive dynamic more visible than before.

How has the relationship between fintech companies and banks evolved from disruption to collaboration?

Financial services run on a long and complex value chain, from infrastructure and compliance to customer interface and product design. What we are seeing now is greater clarity about who wants to own which part of that chain.

Banks increasingly make deliberate choices. In some areas, they want to build and control capabilities in-house, especially where scale, balance sheet strength, or regulatory expertise matter. In other areas, they prefer to rely on specialized partners and integrate external solutions.

This is where fintechs come in. At certain points of the value chain, they are natural collaborators. At others, especially in customer-facing or high-margin segments, they compete directly.

This dynamic is not unique to finance. In many service ecosystems today, firms can be partners in one layer and competitors in another. The same logic now defines the evolving relationship between banks and fintechs, and we can see that this is a mutually beneficial relationship for them.