E.SUN Bank and IBM create an AI governance system for safe AI deployment at the bank

E.SUN Bank together with IBM is working on an AI governance framework that should determine how artificial intelligence can be used within the bank. For the financial sector, this is no longer an experiment, but an attempt to turn the rapid growth of AI tools into a managed and verifiable system.

Why Control Is Needed

Banks are already applying AI not in pilots, but in daily processes: to check suspicious transactions, assess credit applications, and process customer inquiries. The wider these scenarios, the higher the cost of errors. If a model rejects an application without a clear explanation, misses fraud, or gives an employee incorrect guidance, the problem quickly becomes not technical, but regulatory and reputational.

So now the focus is shifting from the question "where to implement AI" to "how to keep it under control." The partnership between E.SUN Bank and IBM reflects exactly this shift.

Financial organizations need not only a model with good metrics, but also a set of rules that describes its entire lifecycle: from data selection and testing to monitoring after launch. In a banking environment, it's not enough to simply prove that the system works on average. You need to understand in what cases it can fail, who has the right to change parameters, how decisions are recorded, and when human intervention is mandatory.

What Will Be Included in the Framework

Although the project details are not disclosed in the brief description, the idea of AI governance usually comes down to several mandatory layers of control. A bank must not only allow the use of AI, but also determine its application boundaries in advance. This is especially important where decisions affect money, personal data, and customer access to services. In practice, such a framework usually includes policy, approval processes, quality control, and distribution of responsibility between business, IT, and compliance.

• criteria for which products AI is acceptable and where only humans are needed
• requirements for data quality, model testing, and bias checking
• audit of decisions, change logging, and a clear line of responsibility
• monitoring procedures after launch, including triggers for review or model shutdown

For IBM, such projects logically fit into the corporate agenda around trustworthy AI and risk management. For E.SUN Bank, it's a way to not slow down the implementation of new tools, but rather make it predictable. When rules are described in advance, it's easier for the bank to scale AI across divisions: anti-fraud, scoring, support, internal operations. Without such a common framework, each team starts inventing their own order, and with it come inconsistencies, duplicate controls, and legal risks.

Why Banks Are in a Hurry

The financial sector has special motivation here. Banks operate under strict supervision, store sensitive data, and make decisions that directly affect customers. So even beneficial AI cannot simply be embedded in a process and expected to work like ordinary software.

Models change with data, can degrade over time, and sometimes behave unpredictably on rare cases. The more actively the market tests generative AI and customer service automation, the stronger the demand for unified control rules. This leads to a broader conclusion for the industry: the next wave of competition will be driven not only by the quality of the models themselves, but also by the maturity of their management.

The banks that will win are those that can simultaneously accelerate processes and prove to regulators, auditors, and customers that AI is used transparently. In this sense, governance becomes as important a part of the infrastructure as cybersecurity, access management, or backup of critical systems. Without it, AI remains a set of scattered experiments.

What This Means

The E.SUN Bank and IBM project shows that for banks, the main issue is no longer "do we need AI," but "what rules does it work by." If such frameworks become the standard, the market will quickly move from point pilots to mass, but more controlled implementation of AI in financial services.