Banks Face Governance Hurdle as AI Moves From Testing to Production

The financial services sector is transitioning AI from experimental phases into broader operational use, with banks, capital markets firms, mortgage lenders and insurers all moving toward larger-scale implementations. This shift brings a new set of demands: managing AI responsibly across the organization while adhering to strict regulatory, security and compliance standards. The AWS Financial Services Symposium in May underscored this inflection point in how the industry approaches artificial intelligence.

Scott Mullins, former general manager of AWS Worldwide Financial Services at Amazon Web Services Inc., emphasized the value of testing before full commitment. "You can spin this stuff up really quickly; you can experiment with it," he said during the symposium. "You can show that it actually achieves the things that you want to achieve within specific policy guidelines and frameworks, and you can actually evaluate whether that governance is holding."

From AI pilots to regulated production

Transitioning AI from experimentation to production-grade systems requires financial institutions to tackle a fundamentally different challenge set. Organizations must go beyond proving that the technology functions as intended; they need to build the governance structures and operational safeguards required to run AI across regulated business processes.

Paul Nashawaty, practice lead and principal analyst for application development at theCUBE Research, highlighted the gap between AI adoption and operational readiness. "Our AppDev research shows that 92% of organizations now integrate AI into at least one stage of the software delivery lifecycle, yet only about one-third have established the governance, security, and operational maturity required for production-scale AI," he said. "In banking, capital markets and insurance, the gap is even more pronounced because compliance, explainability, and customer trust are non-negotiable. The next phase of AI adoption will be defined less by model performance and more by platforms that embed policy controls, auditability and security into AI-enabled workflows while improving customer experiences and operational efficiency."

Infosys Ltd. is building solutions to address these requirements through its AI platform, which incorporates safeguards and controls intended to enable financial institutions to deploy AI while managing risk. Rajaram R.K., vice president and head of sales for financial services at Infosys, described the platform's capabilities: "It has the securities [and] it has the guardrails. It has the token economics … all the capabilities that an organization would need."

Trust, governance and resilience become the real AI architecture

Financial institutions operate under stringent regulatory requirements that make security, auditability and recovery capabilities non-negotiable. As banks, insurers and capital markets firms scale their AI deployments, these requirements gain additional weight, especially when automated systems begin making or influencing business decisions. Notably, financial institutions are among the sectors leading overall AI adoption despite these constraints.

Nashawaty emphasized the importance of treating AI as a regulated business function rather than merely a technology project. "The organizations that treat AI as a regulated business capability, not simply a technology initiative, will be best positioned to accelerate innovation while maintaining resilience and meeting evolving regulatory expectations," he said.

Implementing established risk-management practices within AI systems requires organizations to examine how automated systems access data, make decisions and operate within existing policies. The central tension lies in extending governance and operational controls to AI-enabled workflows while preserving the efficiency and innovation benefits those systems are designed to deliver.

Customer experience and employee productivity move AI into daily workflows

Practical applications for AI are emerging across customer service and employee productivity functions at financial institutions. At New York Life Insurance Co., automating routine manual tasks is freeing employees to focus on personalized service and proactive customer engagement. Kristina Welke, vice president at New York Life, explained the strategy: "We really felt there was a lot of opportunity to support them, remove the manual work they do every day, enable them to provide that consultative support, be proactive versus reactive."

Fifth Third Bank is deploying conversational AI to enhance customer interactions and lower average call-handling time. The bank relies on Amazon Connect to power its customer-service operations. Kevin Anderson, product manager on the AI transformation team at Fifth Third Bank, described the foundation: "Amazon Connect really lays the foundation as it solves for not only messaging but voice as well, allows for AI and automation."

Cloud modernization becomes the foundation for AI deployment

Scaling AI across financial institutions depends significantly on modernizing the infrastructure that handles data-intensive operations. Cloud platforms offer the flexibility to introduce new capabilities while maintaining connections to legacy systems, making infrastructure modernization a critical component of AI expansion strategies.

Mortgage lender PennyMac Financial Services recently expanded its partnership with Amazon Web Services to accelerate modernization of its proprietary mortgage servicing platform, Plaisse. The expanded agreement includes plans to deploy enterprise-grade AI capabilities and continue developing the platform.

As financial institutions upgrade their infrastructure, they also require solutions for moving and managing data without interrupting existing operations. NetApp Inc. is providing storage and data-management capabilities designed to support migration, resilience and performance in AWS environments. Jacob Lim, cloud solutions architect at NetApp, outlined the approach: "We create value for customers by making it simple to migrate and run their data in AWS by combining proven NetApp ONTAP capabilities, such as space-efficient snapshots, near-instant data cloning capabilities with minimal storage overhead and storage performance. Then, there is also the highly efficient data replication and ransomware detection engine natively built into the operating system directly. Combined with AWS scale and elasticity, customers gained resilience, performance and operational simplicity in the cloud."

Scaling AI in financial services brings new challenges

Managing the operational complexity of AI deployments is emerging as a critical challenge for financial institutions incorporating AI into daily operations. According to PagerDuty's 2026 State of AI-First Operations report, 59% of surveyed banks were actively incorporating AI into their operations, while another 34% were discussing implementation. As adoption accelerates, institutions must determine how to govern these systems, maintain data quality and integrate AI into existing workflows.

Todd McNabb, chief revenue officer of PagerDuty Inc., characterized the evolving focus: "I think you're going to see AI really become: How do we govern, how do we use [and] how do we make sure the data's clean? How does it become [about] helping us be more efficient in what we're doing in day-in and day-out operations?"

Technology providers are responding to demand for faster enterprise AI deployment. AWS announced a $1 billion investment in forward-deployed engineering intended to help customers develop and deploy agentic AI applications more rapidly. For financial institutions, however, success will be measured not only by deployment speed but also by how effectively AI operates within existing systems, regulatory requirements and business processes.