Priya Jaiswal brings a wealth of knowledge to the table regarding the digital transformation of the banking sector. As a recognized authority in portfolio management and international business trends, she understands that the transition from AI experimentation to full-scale production is the ultimate litmus test for modern financial institutions. Today, we explore how leading organizations are navigating the delicate balance between rapid technological deployment and the steadfast maintenance of customer trust and regulatory compliance.
We begin our discussion by examining the tension between operational efficiency and the necessity of human oversight in automated systems. Our conversation covers the vital role of collaborative fraud intelligence networks and how “tech-to-touch” strategies are revolutionizing financial inclusion for non-traditional borrowers. Finally, we dive into the evolving landscape of cybersecurity, focusing on the defense against sophisticated deepfake threats and the critical importance of upskilling the workforce to ensure humans remain at the heart of technological progress.
Moving AI into production often creates a tension between operational speed and oversight. How do you embed human oversight into data practices to ensure that automated decisions remain fair and transparent?
Integrating AI into a legacy environment is essentially a control problem where we must prioritize fairness and transparency above all else. At RCBC, we ensure that human oversight is not an afterthought but is deeply embedded into our data practices from the very start of the development cycle. We are closely monitoring regulatory frameworks, such as the standards recently established in Vietnam, to serve as a blueprint for responsible deployment across the region. By establishing clear rules for data use and accountability, we ensure that as our models move from pilot phases into production, they remain consistent with the high level of trust our customers demand. Ultimately, the goal is to create a system where technology handles the scale, but human officers maintain the final say on the ethical implications of data use.
The rise of sophisticated, cross-border cyber threats suggests that individual banks cannot fight fraud in isolation. What is the significance of participating in shared networks like the Fraud Intelligence Data Sharing (FIDS) system?
The reality of modern cybercrime is that threats are no longer confined within the walls of a single institution; they are fluid, sophisticated, and frequently cross organizational boundaries. To combat this, RCBC joined the Fraud Intelligence Data Sharing (FIDS) Network as an initial member, a collaborative effort launched by CIBI Information and FinTech Alliance.PH. This network allows us to contribute verified reports of confirmed or suspected fraud to a shared database, creating a collective shield that benefits the entire industry. When a new applicant’s details have been flagged by another member, we receive a real-time alert, allowing us to identify risks that would have otherwise remained hidden in a fragmented system. This level of cooperation is the only way to effectively counter criminal networks that thrive on the gaps between different financial records.
Many traditional banking models struggle to serve individuals without standard credit histories. How is the shift toward “tech-to-touch” experiences and alternative data changing the landscape for first-time borrowers?
We have transitioned toward a “tech-to-touch” philosophy that utilizes AI to solve the historical problem of financial exclusion for self-employed professionals and first-time borrowers. By moving away from manual processing to an in-house credit-decisioning engine, we can now offer same-day approvals for preferred segments while maintaining rigorous risk controls. Our AI models are specifically designed to analyze alternative data alongside traditional credit scores to find creditworthiness where conventional evaluation methods see a blank slate. This automation removes significant operational bottlenecks at the back end, which in turn frees our human staff to focus on high-value interactions. This allows our employees to provide the empathy, context-specific problem-solving, and personalized financial advice that a machine simply cannot replicate.
As fraudsters adopt tools like deepfake voices and cloned websites, the old static security rules are becoming obsolete. How should institutions adapt their defensive strategies to counter these real-time behavioral anomalies?
As criminals weaponize technology through deepfake voices, cloned websites, and spoofed SMS, the banking industry has to move beyond static, rules-based systems that are easily bypassed. We have shifted our focus toward machine learning tools that ingest large datasets to detect behavioral deviations in real time, which is essential for catching social engineering threats before they affect clients. A practical example of our commitment to security is the strict single-device login policy implemented for the RCBC Pulz mobile app, which is designed to stop unauthorized access in its tracks. Furthermore, our intelligence systems are now capable of identifying rapid fund movements across mule account networks. This allows us to adjust our controls and act against criminals immediately while ensuring that legitimate users experience as little disruption as possible.
The human element is often cited as the biggest hurdle in digital transformation. How do programs like the AIVenger initiative bridge the gap between complex AI capabilities and the everyday workflows of banking officers?
The most challenging aspect of digital transformation is not the software itself, but ensuring that the workforce can effectively collaborate with the technology rather than fearing it. We tackle this head-on through our AIVenger program and regular hackathons, which empower our officers to become AI-certified and proficient in writing tailored prompts for internal problems. Rather than relying solely on third-party systems, we have invested heavily in our own in-house data scientists who build, train, and refine proprietary models based on real-world performance. This internal capability ensures that our staff views AI as a tool for augmentation, allowing them to apply human judgment to complex tasks. By constantly upskilling our team, we ensure they are equipped to handle the judgement-heavy workflows that require deep context and empathy.
What is your forecast for the future of AI in the banking sector?
AI will move from being a specialized tool for fraud or lending to becoming the invisible backbone of every single customer interaction. We will see a significant shift where proprietary models, developed and refined by in-house data scientists, take precedence over generic third-party solutions to ensure better alignment with specific local market needs. The “tech-to-touch” model will become the industry standard, where back-end complexity is entirely automated to allow for a return to highly personalized, relationship-based banking. Ultimately, the winners in this space will be the institutions that can use AI to identify and stop threats like mule account networks in milliseconds, while simultaneously making the banking experience feel more human than ever before. Real-time behavioral analysis will replace passwords and static checks, creating a seamless environment where security is robust yet entirely friction-free for the user.
