AI helps pave way for autonomous banking

In the early stages of digital transformation, banks focused on digitizing individual processes. While this improved efficiency, it did not create a seamless customer journey. The problem is that various banking functions often operate with separate systems. Customer data may reside in one system, credit information in another, and transaction records in a third. These disconnected systems make it difficult for banks to gain a comprehensive customer view, or "Customer 360," which is crucial for informed decision-making.

To address this issue, banks must connect data across their organization. This shift leads to Cognitive Banking, where each function uses its intelligence to understand customer needs and make better decisions. By integrating a complete customer view, including identification, payment history, transaction behavior, and product requirements, banks can enhance customer service, risk management, and sales strategies.

However, merely integrating data is not enough to reap the full benefits of AI in banking. AI must be used judiciously, with human oversight for critical decisions like credit approvals. The level of automation depends on the task, associated risks, and the bank's policies and controls. As banks adopt more AI agents, monitoring their activities becomes increasingly important to ensure transparency and accountability.

In Vietnam, banks must consider data sovereignty when adopting AI. They should determine where their data is stored and processed, especially when using AI models from external providers. Banks have several deployment options, such as hosting in-house, using private or public clouds, or adopting a hybrid model. The choice depends on factors like existing infrastructure, business needs, and regulatory requirements.

Banks must also decide between building their AI solutions in-house or purchasing them from the market. This decision involves weighing the costs and benefits of each approach. Regardless of the path chosen, banks must ensure proper monitoring, clear rules, and traceability of AI's actions to maintain control and uphold data sovereignty.