The Bangko Sentral ng Pilipinas (BSP) is tightening AI governance. In June 2025, the central bank released proposed guidelines requiring all supervised financial institutions to establish a model risk management framework — covering validation, implementation, and ongoing monitoring. Flawed models, the BSP warned, may lead to “financial losses, poor strategies, and reputational damage.”
This regulatory push comes as AI adoption in Philippine banking remains nascent. Many institutions are still running pilots. The industry is shifting from experimentation to accountable, scalable deployment — and BSP supervision is accelerating the timeline.
24/7 customer support: Always on assistance at scale
BPI has led the way with BEA Chat, a 24/7 digital assistant accessible via the bank‘s website and Facebook page. It delivers responses in as fast as three seconds, handling product inquiries, self service options, and seamless escalation to live agents.
BPI’s chief customer officer noted, “BEA Chat allows us to be present for our customers 24/7, whether they‘re at home, at work, or even overseas.” The lesson is clear: customers expect immediate answers at any hour. Traditional contact centers are linear — one agent, one conversation. AI agents handle thousands of simultaneous interactions without shift constraints or attrition.
Instant loan inquiries: From days to minutes
UnionBank, which aims to reach 20 million customers in 2026, will soon launch RIA — a voice AI customer service agent designed to reduce wait times and enable calls at any hour. RCBC Credit has already validated the loan inquiry use case: its AI assistant, Erica, handles over 600,000 interactions annually, automating routine eligibility and documentation questions while freeing human agents for higher value tasks.
Instant loan inquiry handling offers three advantages: lower abandonment (customers get answers immediately), reduced cost per interaction (AI is a fraction of human handled calls), and higher conversion rates (interested customers can be routed directly to loan officers while engagement is high).
Real time fraud s: Protecting customers as threats evolve
The Philippines has one of the fastest growing digital payment markets in Southeast Asia, and with growth comes fraud. According to a Philippine Institute for Development Studies report, digital fraud is a key driver of distrust among Filipino consumers.
Regulators are responding. BSP mandated AFASA (Anti Financial Account Scamming Act) compliance requires all supervised institutions to upgrade fraud detection systems by June 2026 — including real time monitoring, AI detection, and multi factor authentication. Traditional batch processed s arrive too late. AI powered systems monitor transactions in real time, flag anomalies instantly, and notify customers via chat or SMS. UnionBank has already been exploring AI for suspicious pattern identification and fraud mitigation.
Moving from pilot to scale
BSP’s coming wave of regulations — AFASA by June 2026, model risk management guidelines under consultation — means banks can no longer afford fragmented, isolated pilots. The three use cases above share common requirements: omnichannel coverage, real time intent recognition, deep integration with core banking systems, and automated audit trails for compliance.
Instadesk ChatBot is built for this shift. It enables Philippine banks to deploy AI agents that work across channels, understand intent instantly, integrate with loan and fraud systems, and maintain full auditability. For banks ready to move beyond pilots, the path forward is clear.
