Banks Don’t Want Another AI Tool; They Want an Agent Factory

The primary obstacle for banks integrating AI is not the technology itself, but rather the quality of their data. Institutions with poorly organized data often require more complex and expensive AI models to achieve desired outcomes. Ant Digital Technologies addresses this by helping banks structure their data first, which can enable the use of smaller, more cost-effective models and improve project ownership beyond just IT departments.

The application of AI in banking is evolving from single-task tools to multi-functional "agent factories." This shift reflects a desire for comprehensive solutions that can manage various banking operations, such as underwriting, KYC, and fraud detection, which typically rely on distinct systems. Ant Digital Technologies' "Commercial Agent Super Factory" and Agentar 2.0 platform, with its role-based templates and skill modules, illustrate this approach.

Agentic AI, which can act on a customer's behalf, introduces new accountability challenges. While earlier technology required direct human operation, AI agents function as extensions of their users, meaning the user remains accountable for the AI's actions. This necessitates organizational adjustments to processes, controls, and audit practices to align with AI investments and avoid disappointment.

Smaller, established financial institutions, such as thrift and rural banks in the Philippines, may find an advantage in AI adoption despite legacy infrastructure challenges. With fewer systems to untangle compared to larger traditional banks, they can potentially move faster once committed. These banks are already using AI to streamline back-office tasks, improve credit decisions, and enhance fraud detection, contributing to profit growth in the sector.