The NCUA Board met on June 24 and confirmed that Phase 1 of its Deregulation Project has wrapped up 31 proposed rule changes, with hundreds of comments still under review. Phase 2 will focus on complex policy and operational initiatives, where technology governance, third-party risk, and AI oversight live. Proposals are forthcoming.

Why it matters for your CU: Credit unions already building documented AI governance can shape Phase 2 rules through the comment process. Those that aren't will implement rules written without their input.

FinCEN, NCUA, OCC, the Federal Reserve, and FDIC jointly proposed customer identification requirements for permitted payment stablecoin issuers under the GENIUS Act on June 18. NCUA is specifically named as a supervisor for any stablecoin issuer operating as a CU subsidiary. Comments are open 60 days from June 22.

Why it matters for your CU: If your credit union is exploring digital product expansion or fintech partnerships involving digital assets, this defines your compliance framework. The comment window closes around July 22.

A PYMNTS panel featuring executives from Velera and Rubrik concluded that the biggest AI advantage for credit unions won't come from being first to deploy. It will come from redesigning workflows around better outcomes. Panelists flagged the shift from automation to decision support as the moment where member trust becomes a real competitive differentiator.

Why it matters for your CU: Speed is not the winning strategy. Credit unions that redesign workflows around member outcomes, with humans at the right decision points, hold the trust advantage as agentic AI matures.

Walk into any credit union technology meeting right now, and you will hear the same thing. We have Copilot. Our team loves it. We are covered on AI.

And look, Copilot is good. If your loan officers are using it to draft member emails, pull up policy documents, or summarize a meeting, it is doing its job. Nobody is saying throw it out.

But sit with a funding team for a day and watch what actually happens when an indirect auto packet comes in.

The dealer sends over a package. Something is missing; the insurance document is not there. Someone on your team notices, pulls up the file, emails the dealer, and waits. The dealer sends something back. Someone opens it, checks it against the requirements, and either kicks it back again or adds it to the file. Then the file goes to underwriting. Then someone checks it against your lending policy. Then it gets funded in the core.

Copilot does not do any of that. It helps the person sitting at the desk think faster. The desk is still there. The emails still go out manually. The core still gets updated by a human. The handoffs are all still happening; they are just happening a little more smoothly.

That is the gap. Copilot is a thinking tool. Your lending operation needs an acting tool.

Agentic AI does not help your team work through the packet faster. It works through the packet. It identifies the missing document, sends the follow-up, checks what comes back, applies your lending policy, and moves the file forward inside your LOS, your core, your document system, all at once.

Your loan officers do not disappear. They stop doing the parts of the job that were never really their job. The follow-up emails. The re-entry. The waiting. They focus on the member and the decision. That is what they were hired for.

The credit unions figuring this out are not replacing Copilot. They are using both. One to help their team think, one to help their institution act.

Source: NCUA Q1 2026 State-level Credit Union Data Report

🔔Subscribe here