Karen C. Harbin, President and CEO of Commonwealth Credit Union in Frankfort, Kentucky, joins host Rodney Hood to share how watching her mother keep books at home led to a 40-year career at one credit union, which she has grown from $897 million to about $3 billion in assets since 2012. Karen explains why Commonwealth went live with Zest AI underwriting in 2021 to cope with departing underwriters and say yes to more members, how her team keeps humans in the loop, and what a credit card pullback taught them. She and Rodney also cover fraud and cybersecurity, why the efficiency ratio is the metric she watches most, and her advice for new NCUA Chairman John Crews on the interest rate cap and loan maturity limits.

Chapters

00:00 Welcome Karen C. Harbin of Commonwealth Credit Union
01:19 From green bar paper to a $3 billion credit union
03:32 Why Commonwealth went live with Zest in 2021
06:54 Primary financial institution isn't the goal anymore
08:47 Human in the loop and the credit card lesson
11:46 How AI is changing the workforce
14:22 The problem that led to Zest
16:08 Fraud, Backroads to Better Money and a cybercrime certificate
20:36 The efficiency ratio
22:10 What the chair of America's Credit Unions reads
24:35 Advice for NCUA Chairman John Crews: rate cap and loan maturities
27:19 Lightning round: personal finance, Warren Buffett and Unreasonable Hospitality

Takeaways

- Commonwealth adopted Zest AI underwriting in 2021 when tenured underwriters were leaving during the Great Resignation.
- Automated underwriting lets the credit union say yes to more members while delinquencies and charge-offs stay below industry standard.
- Humans stay in the loop: the team sets loan policies, compares decisions with the model and reviews portfolio trends.
- When credit card delinquency rose, Commonwealth found it had gone above Zest's recommended score threshold and pulled back.
- Karen's key metric is the efficiency ratio paired with operating expenses; Commonwealth is under 60%.

Connect with the Guest

Karen C. Harbin LinkedIn: https://www.linkedin.com/in/karen-harbin-cpa-575a0514/
Company Website: https://www.ccuky.org

Sponsor

The Origination Podcast is brought to you by Zest AI, which provides AI-driven underwriting, fraud detection, and lending intelligence for credit unions and banks. Learn more at zest.ai.

Michael R. Wilson, President and CEO of Members 1st Federal Credit Union, joins Curtis Wagner to explain how he rebuilt trust inside the credit union after a five-year succession period, and why that trust is what lets Members 1st move fast on AI. Mike shares how associates picked nearly 1,000 tasks to automate, how his board ran a strategic retreat with six AI board members, and why he sat in on the AI underwriting RFP himself.

For credit union and bank leaders weighing AI in lending, Mike makes the case that CEOs own third-party due diligence. If a vendor can't show you how its decisioning model works, you can't explain it to your regulator, and "that's proprietary" is not an answer.

Chapters

00:00 Welcome: Curtis Wagner and Mike Wilson of Members 1st
00:47 From public education to the governor's cabinet to credit union CEO
02:29 Taking a chance on people who deserve a chance
03:33 Calculated risk, owning failure and no surprises for the board
05:16 Repairing a culture after a five-year succession
06:31 Associate-driven AI: nearly 1,000 tasks to automate
09:43 AI governance and a board retreat run by AI board members
13:27 Choosing an AI underwriting partner you can explain to a regulator
16:19 Crossing $10 billion and third-party vendor due diligence
18:41 Lightning round: 90s hip hop and wine pairings

Takeaways

- Members 1st made healing its culture priority number one after a five-year succession period, and protecting it priority one A.
- Associates named, ranked and prioritized the nearly 1,000 mundane tasks Members 1st is about to automate, which drives excitement instead of fear.
- AI governance should be built before deployment, around the regulatory non-negotiables you already know, because regulation of AI in lending is coming.
- Mike chose Zest AI because it was the only finalist willing to show transparently how its decisioning model works.
- CEOs own third-party vendor due diligence, from model explainability to SOC 1, SOC 2 and cybersecurity, and boards should hold them accountable.

Connect with the Guest

Company Website: https://www.members1st.org

Sponsor

The Origination Podcast is brought to you by Zest AI, which provides AI-driven underwriting, fraud detection, and lending intelligence for credit unions and banks. Learn more at zest.ai.

Joel Swanson, EVP and Chief Lending Officer at VyStar Credit Union, joins host Curtis Wagner of Zest AI to share how a pricing pitch to a credit union CEO changed the way he saw the industry, and how he took over lending right as the post-pandemic credit cycle turned. Joel explains why so many 700+ credit scores turned out to be risky, what VyStar found when it checked whether its AI underwriting model was to blame, and how it has handled 22% more applications with 15% less staff without layoffs. He and Curtis also cover explainability and regulators, VyStar's head-and-heart approach to community, his contrarian view on rates, and why assets are the most overrated metric in credit unions.

Chapters

00:00 Welcome Joel Swanson of VyStar Credit Union
00:30 From Alaska to a CEO who asked the opposite question
02:19 The pivot that almost broke him, and the stupid question
05:01 4,000 miles to Jacksonville
07:08 Taking the lending chair as the fake economy ended
09:37 The 700 credit scores that should never have gotten a loan
11:50 Smarter, faster, cheaper and AI without layoffs
16:08 Explainability, regulators and model validation
19:01 Head and heart at VyStar
24:25 Deposits, fraud, rates and the metrics that matter

Takeaways

- Pandemic-era credit scores overstated many borrowers' ability to repay, and the losses showed up at the end of 2023.
- Some of the riskiest loans went to borrowers with 700+ credit scores, and a few large misses outweigh many small ones.
- When VyStar checked whether its AI model was behind rising charge offs, it found the model was its most effective data point.
- VyStar handled 22% more applications with 15% less staff through attrition, which reduced pushback on later AI projects.
- A credit union has to balance head and heart: smart business decisions made for the right reasons.

Connect with the Guest

Joel Swanson LinkedIn: https://www.linkedin.com/in/joel-swanson-9a99a41/
Company Website: https://www.vystarcu.org

Sponsor

The Origination Podcast is brought to you by Zest AI, which provides AI-driven underwriting, fraud detection, and lending intelligence for credit unions and banks. Learn more at zest.ai.

Jeremy M. Pinard, Chief Lending Officer at Canvas Credit Union, joins host Rodney Hood, former NCUA Chairman, to share how a would-be Secret Service agent found his way into lending and what it takes for a $5 billion credit union to originate like a $10 billion one. Jeremy explains why SoFi-style instant decisions are now the benchmark, how Canvas is working toward 70–90% auto-decisioning, and why success means twice the volume with the same team. He and Rodney also dig into "no margin, no mission," the $10 billion regulatory threshold, the metrics and economic signals he watches, and the leadership habits he uses to build a bench.

Chapters

00:00 Welcome and introducing Jeremy Pinard
01:20 From the Secret Service to a bank in Colorado
02:31 What shaped his leadership, and a setback that taught him most
05:23 Originating 24/7 and the push to 90% auto-decisioning
07:53 The cost of "we'll get back to you shortly"
10:08 Buzzwords, metrics and the economic signals that matter
12:53 A 2027 prediction and the regulatory wish list
15:37 No margin, no mission
18:38 Recruiting talent and the one-on-one question
22:12 Warren Buffett, staying current and AI at home

Takeaways

- Credit unions now compete with apps like SoFi that decide in seconds, so 24/7 instant origination is the new baseline.
- Canvas Credit Union is targeting 70–90% auto-decisioning so staff can focus on members who actually need a person.
- The real measure of lending automation is twice the volume with the same number of people.
- "No margin, no mission": pricing for risk and earning a healthy ROA fund both members and the talent that serves them.
- Leaders earn trust by owning mistakes, correcting them openly, and developing successors.

Connect with the Guest

Jeremy M. Pinard LinkedIn: https://www.linkedin.com/in/jeremy-m-pinard/
Company Website: https://www.canvas.org

Sponsor

The Origination Podcast is brought to you by Zest AI, which provides AI-driven underwriting, fraud detection, and lending intelligence for credit unions and banks. Learn more at zest.ai.