Federal regulators just sent community lenders their clearest signal of 2025: more room to move, heavier expectations for control.

The FDIC proposed lowering the Community Bank Leverage Ratio from 9% to 8%.
The OCC cut back AML data collection, citing low risk across community institutions.

Less paperwork. More flexibility.

At the same time, European Central Bank (ECB) supervisors released their fourth year of AI on-site reviews and found the same underlying issue across banks: they can’t justify model decisions because the data coming in isn’t validated.

That’s the real story: Regulators are easing the load - while expecting institutions to own the quality of everything entering the system.

Here are three shifts showing where 2026 is heading.

AI ROI Is Surging -
But Only for Institutions With Clean Intake Data

🔍 What’s going on:
Three of the biggest U.S. banks just crossed a milestone: AI officially moved out of “pilot mode”.

• Wells Fargo promoted AI into the C-suite and trained 90,000 employees.

• Citi freed 100,000 developer hours per week.

• JPMorgan is letting staff draft performance reviews with an internal LLM.

Across corporate America, 90% of CFOs now report strong ROI from generative AI, up from just 27% last year.

The signal is clear: Corporations that integrate AI into workflows are seeing returns.
Banks are beginning to follow, but with stricter expectations around explainability and data governance.

💡 Why it Matters

Banks don't have a model accuracy problem.
They have a data-readiness problem — which is completely normal in early AI adoption.

ECB supervisors found that 88% of significant institutions now use AI for fraud detection, compliance, and operational efficiency - yet many still can't explain why their models make specific decisions. The gap isn't model sophistication. It's data governance at the point of entry.

If an AI model can't act on the application data alone, the issue is almost always upstream - before the file ever reached the model.

⚡ Action steps

Review 10 recent approvals. Note where staff had to step in - clarifying data, resolving inconsistencies, or verifying identity manually. More than a few interventions per file? Your intake process needs strengthening before AI can deliver its full value.

Fraud Is Moving Faster Than Institutional Responses -And It's Hitting You at Intake

🔍 What’s going on:
Recent fraud resiliency research found most institutions take days to respond to confirmed fraud - not from slowness, but from unclear ownership and misaligned definitions across teams.

The key finding: Fraud doesn't attack your core systems. It attacks the gaps between them.

Security experts are blunt about the threat: "Static data is dangerous." Fraudsters aren't failing KYC anymore. They're succeeding through VoIP numbers, aged email accounts, synthetic identities built from real fragments, document metadata that contradicts the story in the file, and AI-generated documents that pass visual inspection.

💡 Why it Matters
Regulators reduced AML reporting burdens and eased capital requirements — but raised expectations for fraud governance and model oversight.

The expectation is clear: institutions must own their controls earlier in the process. Fraud that passes initial checks becomes everyone's problem — underwriting, compliance, and every AI model downstream.

European supervisors have been explicit: poor inputs produce unreliable results, no matter the model. U.S. regulators are heading the same direction — with community lenders getting time to strengthen foundations first.

⚡ Action Step: 
Write your fraud workflow in one line:
Detection → Decision → Customer notification → Remediation

Now ask yourself: Who owns each step?
If the answer takes more than 30 seconds, coordination gaps - not technology - are creating fraud surface area.

Borrowers Aren’t Judging You on Approval -
They’re Judging You on Intelligence

🔍 What’s going on:
Small businesses aren’t worried about getting approved anymore. 83% of SMBs believe they’d qualify for credit - the highest confidence level ever recorded.

Credit access has faded into the background. What borrowers pay attention to now is how intelligent their financial partners feel.

They compare experiences across tools they already use: dynamic limits, virtual cards, cash-flow alerts, automated reconciliation, accounting and invoicing integrations, and real-time recommendations tailored to their business.

Payments infrastructure is already shifting: AI agents are beginning to initiate transactions, not just process them. Lending will follow — but only where intake data is structured enough to support it.

💡 Why it matters: 

Borrowers aren't benchmarking you against other community lenders. They're benchmarking the experience against software.

Approval speed isn't the differentiator anymore - the intelligence behind the "yes" is.

But intelligence depends on structure.

The institutions building adaptive credit capabilities - dynamic limits, tailored recommendations, real-time decisioning - all fixed the same thing first.

⚡ Action Step: 
Measure two numbers from your fastest loan last month:

1. Time to complete and verify the intake file

2. Time from clean file to approval

If #1 took more than half of total time, intake - not underwriting - is your competitive bottleneck.

🦊 ZorroFi insight

Every trend from the past month points to the same chokepoint: intake.

That's where risk first appears, workflows stall, AI inherits flawed data, fraud slips through, and borrowers decide if you feel modern or manual.

ZorroFi gives community lenders an intake integrity layer built for this reality: real-time document verification, financial cross-checks, identity and behavioral screening, and anomaly detection before files reach your LOS.

When regulators ease capital but raise governance standards, clean intake is your advantage.
When AI scales but depends on verified data, clean intake is your foundation.
When borrowers expect intelligent speed, clean intake is your differentiation.

Fix intake. Then scale.

💡 Want to Dive Deeper?

Demo: I work with CDFIs, community banks, and credit unions on modernizing lending. Want to learn more?

Subscribe: Get the full breakdown every Thursday in my LinkedIn newsletter -Lending Insights. 

🙌 Help Us Grow

Know someone in lending who’d benefit from these insights? Forward this email — and hit reply if there’s a topic you’d like us to explore next.

📅 Next newsletter drops Tuesday, 12/09.