Most credit unions don’t need a perfect AI roadmap. They need a simple, member-focused plan that delivers progress in 6 months—not 6 years.
Most credit union leaders don’t have a technology problem. They have a momentum problem.
Boards agree digital transformation matters. Vendors pitch AI-powered everything. Members expect the same experience they get from big tech. Yet projects stall for months—sometimes years—because teams are waiting for the perfect core, the perfect roadmap, or the perfect budget cycle.
John Janclaes, President of Nymbus CUSO, summed it up well:
“Progress over perfection. Don’t wait for perfection.”
For credit unions trying to build truly member-centric banking with AI, that mindset is the difference between gaining relevance and slowly losing it.
This post, part of our AI for Credit Unions: Member-Centric Banking series, pulls themes from Janclaes’ conversation on The CUInsight Network and translates them into something every CU leader can use: a practical way to move from AI theory to tangible results without betting the institution.
We’ll look at how to:
- Use AI to win new markets of members without abandoning your core
- Build a culture that actually supports digital and AI initiatives
- Modernize step-by-step instead of waiting for a “big bang” core conversion
- Choose AI use cases that prove value quickly and reduce risk
1. Progress Over Perfection: How Credit Unions Should Approach AI
The most effective AI strategies in credit unions start small, focused, and member-led, not big and theoretical.
Janclaes talks about Nymbus helping credit unions “win new markets of members, enhance processes, and modernize technology.” The order there matters. Technology is last for a reason.
Here’s the thing about AI in credit unions: if you start with tools instead of member problems, you burn time, budget, and political capital.
Start with 3 brutally clear questions
Before any AI initiative, leadership should answer:
-
Which member segment are we trying to serve better or attract?
- Gen Z members new to credit
- Small business owners in your field of membership
- Retirees managing income and health costs
-
What member problem are we trying to reduce or eliminate?
- Long call center wait times
- High friction loan applications
- Confusing digital experiences across channels
-
How will we measure success within 90–180 days?
- Higher digital engagement
- Lower call volume or handle time
- Faster loan decisioning
- More product per member
If you can’t answer those cleanly, you’re not ready to pick an AI solution.
Why “progress over perfection” is a risk strategy, not a slogan
Going incremental with AI sounds safer—and it is—but it’s not about playing small. It’s about learning cheaply.
A realistic pattern for most credit unions:
- Phase 1 (0–6 months): One member-facing AI use case (e.g., virtual assistant for basic member service) plus one internal use case (e.g., AI-assisted fraud alerts review).
- Phase 2 (6–18 months): Expand into AI-based loan decisioning support and proactive member outreach (e.g., next best action).
- Phase 3 (18–36 months): Integrate AI insights across channels so contact center, branch, and digital all see the same predictive signals.
You don’t need a perfect, unified data lake on day one. You need a controlled experiment that proves AI can:
- Remove friction your members feel
- Reduce cost or risk your CFO cares about
- Build confidence among your board and staff
2. Culture First: Building a Credit Union That Can Actually Change
Janclaes spends a lot of time talking about culture and growth mindsets. He’s right to do it. Most AI and digital transformations fail not because the tech didn’t work—but because the organization couldn’t absorb the change.
A credit union that wants meaningful AI in member-centric banking needs three cultural shifts.
1. From “project” to “practice”
Digital and AI can’t live as one-off projects owned by IT. They need to become a cross-functional practice.
Practical steps:
- Stand up a small digital/AI council with leaders from lending, member experience, operations, risk, and IT.
- Give it a clear mandate: prioritize AI use cases, approve experiments, monitor impact.
- Meet monthly and publish a one-page update internally. Transparency drives buy-in.
2. From fear of failure to controlled experimentation
I’ve seen credit unions kill good ideas because the first pilot didn’t hit every KPI. That’s a recipe for stagnation.
A healthier pattern:
- Define “acceptable failure” upfront (e.g., “We’re okay if this pilot doesn’t reduce calls, as long as NPS doesn’t drop and we learn what members actually ask.”)
- Cap pilot scope by time, member segment, or channel.
- Turn learnings into clear decisions: scale, pivot, or sunset.
3. From tech as cost center to tech as growth engine
Janclaes points out that credit unions showed serious tenacity during the pandemic. Many rolled out remote tools, digital signatures, and faster communication under pressure.
That mindset—tech in service of member survival and growth—needs to become permanent. AI investments should be framed in terms of:
- New members reached
- Member retention and lifetime value
- Reduced risk (fraud, credit losses)
- Staff productivity and reduced burnout
When tech is positioned this way, culture stops seeing AI as “robots replacing people” and starts seeing it as “tools that let us serve more members, more personally.”
3. Practical AI Use Cases: Where Credit Unions Should Actually Start
Credit unions don’t need 20 AI initiatives. They need 2–4 that directly support member-centric banking and can be delivered with existing or partner infrastructure.
Here are high-impact starting points that align with what Nymbus and similar providers are helping CUs achieve.
AI for member service: smarter, more human support
AI-driven virtual assistants and chat can reduce friction—if they’re deployed thoughtfully.
Strong first use case:
- 24/7 member service assistant on web and mobile that handles:
- Balance and transaction questions
- Card freezes and travel notices
- Branch/ATM information
- Simple FAQ (routing numbers, hours, basic policies)
What works best:
- Clear handoff to human agents when intent is complex or emotion is high (fraud concerns, loan denials, disputes).
- Training AI on your credit union’s specific language and policies, not generic banking FAQs.
- Using the assistant’s data (top intents, failure points) to refine scripts, IVR menus, and website content.
AI for loan decisioning: faster, fairer, more consistent
AI-supported underwriting, when governed correctly, can help you say “yes” more often without blowing up your risk profile.
Examples of practical use:
- Pre-qualification models that surface members likely to be approved before they apply.
- Risk-based pricing support that suggests ranges but still allows underwriter judgment.
- Exception handling: AI flags borderline files where a human review has the highest impact.
Guardrails credit unions should insist on:
- Transparent models with explainable factors (no black boxes deciding member fate).
- Regular bias audits to ensure protected classes aren’t being adversely impacted.
- Policy that humans make final decisions on edge cases and high-risk profiles.
AI for fraud detection: protecting trust in real time
Fraud is one of the clearest ROI cases for AI in credit unions.
Effective patterns:
- Real-time transaction monitoring with anomaly detection that learns normal behavior for each member.
- Adaptive authentication that adds friction (step-up verification) only when risk scores spike.
- Integrated case management so analysts see AI alerts, evidence, and recommendations in one view.
The member-centric lens here: reduce false positives. Members hate declined legitimate transactions more than almost anything. A good fraud model reduces both fraud and member frustration.
AI for financial wellness: from generic advice to personalized coaching
If there’s one area where credit unions can leapfrog big banks, it’s personalized financial wellness powered by AI.
Examples:
- Spending insights that translate raw data into human language: “Your spending on dining out is up 23% compared to last month.”
- Savings nudges: “You received three similar paychecks in a row. Want to automate $50 from each into your emergency fund?”
- Proactive risk alerts: “You’re on track to overdraft in the next 5 days based on scheduled payments.”
This is member-centric banking in action: using AI to spot patterns and offer help before members ask—without being pushy or salesy.
4. Modernizing Without Blowing Up Your Core
Janclaes talks about modernizing technology while helping credit unions stay relevant in a saturated financial market. The reality many leaders face: legacy cores, limited budgets, and vendor lock-in.
The good news? You don’t have to rip and replace your core to implement meaningful AI.
Think “surround and extend,” not “rip and replace”
A pragmatic architecture approach for most credit unions:
- Keep your existing core stable for system-of-record functions.
- Add API-friendly platforms around it for:
- Digital account opening
- Loan origination
- Member engagement and communications
- AI analytics and decisioning
- Use these modern layers as a sandbox for AI-enabled products and member experiences.
Partners like Nymbus CUSO specialize in this model: they help launch new digital brands, test new markets, and run them on modern infrastructure without forcing you to migrate every legacy product on day one.
Use “digital brands” or niches to test AI-powered offerings
One approach I’ve seen work for CUs:
- Launch a targeted digital brand (for example, a Gen Z-focused mobile-only offering) using modern tech and AI capabilities.
- Limit it to a specific geography or member segment.
- Measure different KPIs from your main brand: digital adoption, cost to serve, product per member.
Lessons from this niche operation can then inform broader changes to your flagship credit union brand—with less risk, less politics, and more data.
5. A 6-Month AI Action Plan for Credit Union Leaders
Theory is nice. Board packets need something more concrete. Here’s a realistic 6‑month plan that applies the “progress over perfection” mindset to AI and digital transformation.
Month 1–2: Alignment and problem selection
- Run a member journey workshop with leaders from lending, operations, and member service.
- Identify the top 3 friction points your members feel most.
- Select one member-facing and one internal AI use case tied to those problems.
- Define success metrics you can measure in 90–180 days.
Month 3–4: Partner and pilot design
- Assess whether your current vendors offer AI capabilities you’re not fully using.
- Where gaps exist, evaluate 2–3 specialized partners (for example, virtual assistants, fraud analytics, or decisioning engines).
- Design limited pilots:
- Clear start and end date
- Target segment or channel
- Metrics, dashboards, and governance
Month 5–6: Pilot, measure, communicate
- Launch pilots with tight feedback loops from front-line staff.
- Monitor impact weekly, not just quarterly.
- Capture member quotes, staff feedback, and metrics—both wins and misses.
- Present results to the board as: what we tried, what we learned, how we’ll scale or adjust.
This is how digital transformation becomes something your organization does, not something it talks about.
Where AI-Driven, Member-Centric Credit Unions Win Next
Credit unions proved during the pandemic that they can adapt quickly when their members need them to. The threat now isn’t just from big banks—it’s from fintechs and tech platforms that are already using AI to personalize offers, anticipate needs, and set expectations for service.
The credit unions that stay relevant over the next five years will do three things well:
- Treat AI as a member-experience tool, not a science project. Every initiative ties back to a specific member problem or opportunity.
- Adopt “progress over perfection” as an operating principle. Small, fast experiments beat long, theoretical roadmaps that never get funded.
- Build cultures that embrace growth and learning. From the boardroom to the call center, people understand that AI extends their capacity to serve—not replaces their purpose.
There’s a better way to approach digital transformation than waiting for the perfect moment or perfect stack. Start with the members you care about most, pick one or two AI-powered improvements you can implement this year, and move. The market won’t slow down, but your credit union can absolutely catch up—and even lead—by choosing progress, not perfection.