Three Significant Changes Banks Must Acknowledge

A community banker recently spent three months trying to convince commercial clients to use bill pay, the feature where the bank cuts a check and mails it, so the customer doesn’t have to.

They weren’t interested. They wanted to write their own check, drive to the post office, buy a stamp and lick an envelope.

That same banker’s board is now being asked to develop a serious artificial intelligence (AI) strategy. Both conversations are happening inside the same institution, sometimes simultaneously. That tension, between the pace of technological change and the customers banks serve, defines community banking right now. Three things just shifted simultaneously that make it more urgent.

The Cores Made Their Move
Jack Henry & Associates announced a Google Cloud integration in September 2022. Fiserv followed with a Snowflake initiative packaged as Data Compass. Fidelity National Information Services (FIS) announced its own Snowflake partnership in late 2025. Three companies. Three separate announcements. One direction.

These are not startups chasing a trend. They are the companies that built batch processing into the foundation of American banking and profited from it for decades. When all three move simultaneously, it signals a competitive reality they can no longer ignore.

Nearly 70% of banks cite lack of integration between systems as their primary technology challenge — up from 60% the prior year. Cloud data sharing addresses that directly. But, accepting the core’s migration timeline means accepting the core’s strategic road map, which is not the same as the bank’s.

The Regulatory Floor Just Disappeared, at the Wrong Moment
In early 2026, federal regulators retired Supervisory Letter SR 11-7 and replaced it with SR/CA 26-2. The original document had governed how banks manage risk in analytical and AI models for over a decade. It was prescriptive and gave compliance officers a clear checklist.

The new guidance, SR/CA 26-2, is principles based. The direction in plain terms: use good judgment, avoid undue risk, be responsible.

That sounds like relief, but it may not be.

Banks are deploying AI tools, large language models, custom agents and automated reporting systems, without asking their regulator a single question. Some deployments involve customer data flowing to third-party servers. Some involve model outputs touching credit or compliance decisions. Whether any of that creates an exam finding under the new framework is still untested.

It won’t stay that way. The first significant AI-related data breach at a bank will trigger rapid guidance and enforcement, the way Synapse Financial Technologies did for sponsor banking. Good compliance officers are already scheduling conversations with examiners outside the annual audit cycle, asking how the examination team interprets AI deployment and third-party data retention under SR/CA 26-2. Everyone else is waiting to see who walks through the minefield first.

Workforce Change Is a Board Question, Not an IT Question
AI will displace a meaningful portion of manual operational work at banks, report pulling, data reconciliation, compliance drafting, routine correspondence, on a timeline closer to 2030 than most boards have planned for.

This is not a technology problem. It is a workforce strategy question. The banks that navigate it well will train employees to move up the value chain, define explicitly where human judgment stays in the loop and treat that preparation as strategy rather than overhead.

The analog customer is not going away. What changes is the capacity of the people serving him, because the operational work that consumed the day gets handled faster, and that time goes somewhere more valuable.

Three Decisions That Belong on the Agenda Now

1. 1. Audit what data the bank actually owns. Most institutions know what reports the core produces. Fewer know what underlying data they own, what it costs to extract or what agreements govern portability. That audit should precede any cloud or AI vendor conversation.
2. 2. Engage the regulator before the exam. Ask specifically how examiners are interpreting SR/CA 26-2 on AI deployment and third-party data handling. The banks doing this now will not be caught off guard when the first enforcement action sets precedent.
3. 3. Treat workforce planning as strategy. Define where AI assists and where humans decide. Get teams trained on current tools. Set that line deliberately, or circumstances will set it instead.

The cores have moved. Regulatory guidance has gone from prescriptive to permissive at exactly the moment when technology risk is highest. The institutions that treat these three shifts as connected — rather than three separate problems for three separate departments — are the ones most likely to come out of this period ahead.