Relational underwriting plus AI helps credit unions approve more members, beat predatory lenders, and turn small-dollar loans into real financial inclusion.
Most credit union leaders know this number by heart: about 40% of Americans can’t handle a $400 emergency without borrowing or selling something.
That single statistic sits at the center of financial inclusion. It also exposes something uncomfortable: traditional credit underwriting, built on scores and backward-looking data, routinely fails the members who need credit unions the most.
Here’s the thing about financial inclusion for credit unions in 2025: you can’t fix it with marketing campaigns or new taglines. You fix it with better decisions. And that’s where AI, relational underwriting, and member-centric design finally start to work together.
This post draws on themes from Seth Brickman, CEO of QCash Financial CUSO, and connects them directly to AI for credit unions—especially for leaders who want to grow, stay relevant, and still live their “people helping people” DNA.
We’ll look at how relational underwriting works, where AI fits, and how you can design member-centric credit products that actually keep people away from predatory lenders.
From Credit Scores to Relationships: What Changes
Relational underwriting replaces a narrow view of “Can this person repay?” with a broader question: “What do we know about this member’s behavior and relationship with our credit union?”
Instead of relying almost entirely on a FICO score, credit unions can:
- Use account history (tenure, deposit patterns, savings behavior)
- Look at transactional signals (payroll stability, bill-payment consistency)
- Consider engagement (use of digital tools, past product performance)
This matters because credit scores often under-represent:
- Young adults without long credit files
- Immigrants and thin-file members
- Women and minorities who statistically face higher denial rates
- Members recovering from medical events or temporary income shocks
Relational underwriting says: “We know this member. Let’s use what we know.”
AI then becomes the engine that can:
- Process thousands of relational variables quickly
- Spot repayment patterns humans would miss
- Generate fair, consistent decisions at scale
That’s the bridge between “we care about financial inclusion” and actually approving more good loans.
Why Predatory Lenders Win — And How AI Can Help You Compete
Payday and predatory lenders don’t win because they’re cheaper. They win because they’re fast, simple, and always say yes (for a price).
When a member has a $400 car repair due tomorrow, they’re not comparing APRs on a spreadsheet. They’re comparing:
- “I can get cash in 5 minutes on my phone” vs.
- “I might get an answer from my credit union in a few days”
If your underwriting is manual and score-only, three things happen:
- Slow decisions – which is a deal-breaker in emergencies.
- More “no” decisions – especially for thin-file or low-score members.
- Broken trust – members stop seeing you as their first call.
AI-enabled relational decisioning can flip this dynamic:
- Instant approvals for small-dollar loans based on member behavior, not just scores
- 24/7 access inside online and mobile banking
- Risk-based pricing tuned by AI models, not guesswork
Seth Brickman often frames it this way: when credit unions say “yes” more often and more quickly, they don’t just make a loan—they intercept that member before a predatory lender does.
AI makes that interception scalable.
Building an AI-Ready Relational Underwriting Model
If you’re serious about AI underwriting, start with the data you already have. Most credit unions are sitting on gold and acting like it’s gravel.
Step 1: Define the member behaviors that signal trust
AI models get better when you’re specific. Instead of a vague “good member,” identify measurable signals like:
- Length of membership (e.g., 2+ years)
- Number of consecutive months with positive balance
- Payroll direct deposit frequency and stability
- History of overdrafts and how quickly they’re cured
- Prior loan performance with your institution
These turn into model features that better reflect actual risk, not just generic credit bureau data.
Step 2: Start with a focused product — like small-dollar loans
Seth’s world at QCash focuses on small-dollar, relationship-based loans that:
- Are fully digital and instant
- Use relational underwriting in seconds
- Aim specifically at emergency and short-term needs
That’s a smart place to start with AI decisioning because:
- Ticket sizes are smaller, so the risk is manageable
- Data feedback loops are faster (you see performance quickly)
- The impact on predatory lending exposure is immediate
You don’t need to turn your entire loan book over to AI on day one. Start narrow, learn fast, then scale.
Step 3: Pair AI with guardrails and policy
AI for credit unions shouldn’t be a black box. You’re still a cooperative, not a hedge fund.
Strong programs:
- Set clear approval/decline boundaries that align with your risk appetite
- Use explainable AI techniques so staff can see why decisions were made
- Review models for fair lending and disparate impact by segment
In other words, AI should augment your risk culture, not replace it.
Designing Member-Centric Credit Journeys with AI
Most credit unions say they’re member-centric. Very few design lending journeys from the member’s emotional state during a crisis.
The reality? Someone facing a $400 emergency is:
- Stressed and time-constrained
- Unsure if they’ll qualify
- Skeptical of hidden fees and fine print
AI and automation give you the tools to design a different experience.
Make the experience feel like a relationship, not a test
Concrete ideas:
- Pre-qualify members proactively. Use AI to identify members likely to qualify for small-dollar loans and show them that option before they’re in crisis.
- Embed offers where members actually are. Inside mobile banking, not on a buried web page. A “Need cash fast?” tile in the app can be powered by AI, only visible when relevant.
- Speak human, not credit jargon. If a member is approved, say: “You’re approved for $X at Y% APR. Funds in your account in 60 seconds.”
Use AI for personalization, not just approval
AI can do more than say yes or no. It can help you:
- Tailor loan sizes and terms to actual cash flow, not guesswork
- Suggest payment dates that match deposit patterns
- Offer nudges: “You’ve nearly repaid this loan; would you like to set up a $20/month emergency savings transfer?”
That’s where AI supports financial wellness, not just credit risk. You’re helping members move from emergency borrowing to resilience.
Serving the Underserved: Diversity, Equity, and Inclusion in Practice
One of Seth Brickman’s strongest points is that vulnerable populations—especially minorities and women—are disproportionately exposed to predatory lending.
AI can either perpetuate that problem or help fix it, depending on how you design and govern it.
Where AI goes wrong
If you train models only on historical approvals and denials, you’re encoding yesterday’s bias into tomorrow’s automated decisions. That’s how:
- Low-income communities keep getting lower approval rates
- Women and minority borrowers receive fewer approvals or worse terms
How to use AI for fairer decisions
Credit unions that align AI with their DEI values focus on:
- Alternative and behavioral data instead of just bureau scores
- Regular bias and outcome testing across race, gender, age, and geography
- Policy overrides that let staff say, “The model declined this, but we know this member and here’s why this should be a yes.”
This is where the cooperative model is a massive advantage. You’re not chasing quarterly EPS. You can afford to treat fairness and inclusion as core success metrics, not PR talking points.
And when underserved members experience a fast “yes” from their credit union after years of “no” elsewhere, they don’t just take a loan—they build a relationship. That relationship is extremely hard for big banks and fintechs to copy.
Innovation, Relevance, and the Credit Union Growth Agenda
Seth talks a lot about staying relevant in a saturated market. I agree with the basic premise: if your credit union isn’t visibly innovating, members assume someone else is doing it better.
AI gives you practical levers to:
- Acquire new members who’ve been ignored by traditional banks
- Retain existing members by saying “yes” when it matters most
- Differentiate with true financial inclusion instead of generic product menus
Here’s a simple test for your next strategic planning session:
If a member’s first $400 emergency in 2026 goes to a payday lender instead of to you, did your digital, member experience, and AI strategies work?
If the honest answer is “no,” then there’s your roadmap.
Start with one member-centric, AI-enabled product—like a QCash-style small-dollar loan. Tighten your relational data, deploy an underwriting model, build the UX, and measure results ruthlessly.
When you see:
- Higher approval rates for thin-file members
- Reduced use of external predatory credit in your community
- Strong repayment and low loss rates
…you’ll know you’re on the right track.
Where Credit Unions Go From Here
Financial inclusion isn’t a side project; it’s the competitive battlefield for credit unions over the next decade. Credit scores alone can’t get you there. AI-powered relational underwriting can.
The takeaway is straightforward:
- Use the data you already have to understand members better
- Apply AI to make faster, fairer, more relational decisions
- Design digital journeys that say “yes” before predatory lenders even enter the picture
If your credit union is serious about member-centric banking, 2026 is not the year to watch from the sidelines. It’s the year to pilot AI-based relational lending, measure its impact on your community, and scale what works.
Members are already telling you what they need—often in the form of quiet, one-time transactions at payday shops and online lenders. The question is whether your AI and your lending strategy will be ready the next time that $400 emergency hits.