By Kishan Sundar, Senior Vice President & CTO, Maveric Systems
Every CIO I speak with is facing the same challenge. The board wants AI rolled out across the business, regulators expect every decision to be transparent and auditable, and the underlying technology was built for a time when systems followed predefined rules. Whether AI initiatives succeed or stall depends on how well banks can bridge the gap between the ambitions and the legacy architecture supporting them.
AI-first banking expands the CIO’s role. Going beyond keeping the systems running, CIOs are accountable for making sure AI-driven decisions are trustworthy, explainable, and defensible. The data reflects the gap: McKinsey’s State of AI research shows that 65% of organisations use generative AI regularly; however, adoption is only about one third at enterprise scale. The bigger challenge is scaling AI while maintaining quality, governance and regulatory compliance.
To make the transition credible, CIOs need to make progress across four closely connected areas: customer engagement, technology modernisation, operational efficiency, and regulatory compliance. Success in one depends on getting the fundamentals right in the others.
Customer engagement and growth
AI is already improving customer onboarding, service, product recommendations and personalised offers. But these initiatives only work when customer information is accurate and consistent across the bank. Many banks operate with fragmented customer data. When different systems maintain different versions of the same customer, AI reflects those inconsistencies. Customers receive conflicting offers, different risk assessments and inconsistent communications depending on which system is serving them.
Software delivery has also become part of the customer experience. As banks introduce more AI-enabled services, development and testing cycles have to become faster without compromising quality. The challenge for CIOs is to ensure that the underlying data and software quality support consistent outcomes.
Infrastructure and application modernisation
Most core banking platforms were designed for transaction processing, reliability and operational stability. They were not built for real-time decisioning, continuous data streaming or AI-driven decision intelligence.
Modernisation is essential, but replacing legacy platforms alone does not solve the problem. If poor data quality and weak governance remain, banks process inaccurate information faster.
CIOs are looking for technology partners who can help institutions modernise incrementally, with measurable outcomes at each stage: faster release cycles, reduced defect leakage, and improved system reliability. Modernisation should improve business performance and not merely replace technology.
Operations transformation
Operations is one of the largest cost centres for most banks, making automation a top priority. AI can improve customer service, lending, fraud detection, payments processing and compliance monitoring. However, automation is only as reliable as the data behind it. If AI is making decisions using incomplete or inaccurate information, errors are repeated at a scale that manual teams cannot easily correct.
Take fraud management as an example. A system that generates large numbers of false positives increases investigation costs while creating unnecessary friction for customers. Efficiency gains come when automation is supported by reliable data, strong testing and effective governance.
Technology partners therefore need to bring more than implementation skills. CIOs need partners who have a clear understanding of banking operations and the practical realities of running AI in production.
Regulatory compliance, resilience and privacy
The conversation around AI in banking has evolved significantly. The focus has shifted from AI adoption to AI governance. Regulators are scrutinising how AI systems are governed, monitored and controlled. As AI becomes embedded in critical functions such as credit underwriting, fraud detection and compliance monitoring, boards and leadership teams are recognising that trust in AI is built on a foundation of high-quality, well-governed data.
That trust quickly erodes when the underlying data cannot be relied upon. A bank cannot explain a lending decision if the model was trained on inconsistent or biased data. Nor can it demonstrate fairness or regulatory compliance without clear oversight of how models are developed, tested and monitored.
Governance should not be treated as a final checkpoint before deployment. It needs to be built into the entire lifecycle, from development and testing to production and ongoing monitoring.
The evolving role of technology partners
As AI capabilities evolve, technology decisions will become more complex. Banks need partners who can help modernise systems in a controlled way, integrate AI into existing environments and establish governance frameworks that keep pace with change. CIOs are looking for partners who understand banking, appreciate regulatory expectations and can deliver measurable business outcomes without introducing unnecessary operational risk.
AI-first banking is also changing what CIOs are accountable for. Success is no longer measured solely by system availability or project delivery. CIOs are increasingly responsible for ensuring that AI-enabled decisions are reliable, explainable and compliant. Banks that build these capabilities into their technology foundations will be better positioned to adopt AI with confidence. They will be able to move faster, manage risk more effectively and deliver better outcomes for customers, regulators and the business alike.