Marc Escapa

Cofounder @ Fuse | AI-native LOS for Credit Unions

[Marc Escapa @escmarc]

June's Flashback. Somewhere between 6 states, 4 stages and an 80s themed party, June took my team to 6 great conferences:

→ ACE26 with Michigan Credit Union League & Affiliates in Grand Rapids
→ NAF with National Automotive Finance Association in Irving
→ Engage with The League of Credit Unions & Affiliates in Orlando
→ Elevate with Maine Credit Union League in Portland
→ Excel with New York Credit Union Association in Saratoga
→ Strategic Growth with America's Credit Unions in Cleveland.

900+ leaders in one room at ACE26 alone.

Every time my team comes back from these events, they talk about the same thing. The people.

Leaders who still know their members by name. People who measure a good year by who they helped into a first car or a first home.

There is something special about how Credit Unions show up for their members and their communities, whether they serve a few thousand or a few hundred thousand.

It is the reason we do this work, and the reason we keep showing up at these events.

Grateful to everyone on our team and at every Credit Union who made June what it was.

[Marc Escapa @escmarc]

This is what the Credit Union of the future looks like to me:

Seventy percent of loan applicants move through without a human touching the file.

Account open, loan approved, funds directed to the bank they named in the application. The whole thing resolved before they've closed the tab.

The remaining thirty percent still get a person. But that underwriter isn't checking items off a list anymore. The system has already done the mundane work. What's left is a real conversation about which product actually helps that member get where they're trying to go.

The 70/30 model puts human attention where it belongs: on the thirty percent who actually need it.

Visit https://t.co/578l44mY8u to see how Credit Unions are getting there.

[Marc Escapa @escmarc]

Most lending software relationships end at go-live.

The contract gets signed, the system gets handed off, and the vendor moves on to the next logo.

Ajay Singh, COO of Drivata Finance, recently told me something different.

A year in, Fuse still feels less like a vendor and more like a partner to them. Questions, feedback, input, months after implementation. In his words, they wouldn't be where they are today without us.

The Handoff Cliff is the default in this industry.

We built Fuse to stay past the handoff, not disappear at it.

[Marc Escapa @escmarc]

I spent a day on the floor with one of our Credit Union's underwriting teams last month.

The senior underwriter, who's been there 16 years, said something I'm still thinking about.

"For the first time in a decade, I have an hour in the afternoon to actually call a member back."

That's the part nobody puts in the case study.

Before Fuse, her day was about 80% manual data entry. Pulling docs from email. Checking the credit report against the application against the pay stub, line by line, for every file. By 4pm she was tired and there was always one file she hadn't gotten to.

After Fuse, the data entry is done before she opens the file, and the clean files are auto-decisioned before they hit her queue. What's left is the work she actually went to school for.

The Grey Zone applications. The conversations with members who got declined and deserve to know why.

That's the shift I care about more than any automation percentage.

What Credit Unions compete on is what happens when a member picks up the phone and someone who actually knows them picks up on the other end. That experience gets crushed when the team is buried in data entry.

The underwriter is still the underwriter. They just get their afternoon back so the part of the job that makes a Credit Union a Credit Union can actually happen.

If your team is losing the human work to the manual work, send me a DM. I'll show you what a week looks like after the data entry is gone.

[Marc Escapa @escmarc]

Ask a legacy vendor to change one thing and the number that comes back is often $20,000. Charlotte Nemec at Canopy Credit Union says that price isn't unusual. It's the standard.

That single number explains a quiet tragedy in our industry. Credit Unions stop improving their own systems, because every fix carries a toll. So the rational move becomes leaving things broken.

I call it the "Change Tax".

We don't price the platform that way, and the reason matters more than the savings: when changes are free, institutions actually keep making them.

The real cost of $20,000 was never the money. It was every improvement nobody bothered to ask for.

[Marc Escapa @escmarc]

Most Credit Unions I work with have verification tools. What they're missing is a verification strategy.

Identity and income verification aren't the same problem. The applicant with a strong credit history and a traceable income source doesn't need the same level of scrutiny as someone who triggers two risk flags on intake. Treating them identically adds friction where it doesn't belong and misses risk where it matters.

The strategy worth building isn't about picking the right vendor. It's about designing the right triggers.

Who moves through with a live photo verification?
Where does a bank connection replace a manual pay stub review?

Get those triggers right and verification becomes an experience calibrated to actual risk.

At Fuse we give Credit Unions access to best-in-class verification providers on a pay-as-you-go basis. No multi-year commitment required to test what actually works for your member base.

Visit https://t.co/BAbaeotPI5 to see how we approach it, or send me a DM.

[Marc Escapa @escmarc]

I started my career as a consultant. The job was to help clients get from where they were to where they wanted to be.

That's still the work.

Every CLO I talk to wants to be where Chime is. They also know they cannot wake up there tomorrow.

The gap is real. The pressure is real. And when a vendor walks in and demos something that looks ten years ahead of where the team is on the floor, the response is the same every time. Analysis paralysis.

So at Fuse we built the path along with the destination.

We meet you where you are. Same workflows. Same policies. Then every two weeks we sit together and pick the next thing. Online ID for a small slice of members. Bank-connected income verification on consumer loans. One automation at a time.

A year in, our clients are running 70% automation. They got there in increments small enough that the team never panicked.

This is the part most vendors skip. They sell you the photo of the summit and hand you a map at base camp.

My job is being the guide still with you on week 38, looking at the next switch to flip.

DM me if your last automation project felt like a map and a goodbye. I'll walk you through what bi-weekly actually looks like.

[Marc Escapa @escmarc]

Canopy Credit Union runs a core built in 1978. When we started, Matthew Utesch expected to spend the project teaching us how it worked.

It went the other way.

In his words: maybe with the exception of our head of IT, you understand XP2 better than we do.

That's the part legacy vendors get backwards. They hand a Credit Union a modern logo and quietly leave the hard part, the old systems you actually run on, as your problem.

A platform isn't modern because it's new. It's modern because it can speak fluently to the core you've had for 40 years.

That's the work.

[Marc Escapa @escmarc]

AI-first. That's the headline on every LOS vendor's homepage now.

Half of them mean they added a document-reading model to the same system they had two years ago. The other half mean they retrained their sales deck.

AI-first at the level a CLO should care about looks like this.

The AI runs the full decision loop. It's a Task Processing AI. The question shifts from "did we read the document?" to "is this task complete?"

The model gets tuned every two weeks. The training data is what your team is doing on the floor right now.

When accuracy improves, the work actually leaves the human queue. That's the test that most platforms fail.

The AI is doing the work and the team is still touching every file. The cycle time hasn't moved.

So when you're evaluating vendors, ask one question. "What percentage of applications never get touched by a human in your average client?" If they can't answer, the AI is decoration. If they can answer and the number is under 50%, you're being sold the same thing in a new wrapper.

Send me a DM with the answer the next vendor gives you. I'll tell you what's behind it.

[Marc Escapa @escmarc]

Here's a problem most Credit Unions are sitting on but rarely say out loud: the core conversion they keep postponing because the cost is terrifying.

And the scary part was never the core itself. It's re-integrating everything bolted onto it. That's the Core Tax.

Charlotte Nemec at Canopy Credit Union is playing it differently. Because we know how to write to her core, work that needs a separate vendor today gets absorbed directly. And when she eventually moves cores, we've already committed to replugging her into the next one.

Infrastructure that outlasts the core turns a migration everyone dreads into a routine swap.

[Marc Escapa @escmarc]

Credit Unions have watched fintechs pull ahead on technology for years.
AI has accelerated that pull in a way nobody saw coming.

A gap that was closing slowly is now widening by the month.

A fintech today can spin up a new loan product, test it with a segment, iterate on the application flow, and have data back in a week.

Most Credit Unions I talk to are still waiting on their LOS vendor to return a support ticket from three months ago.

Meanwhile, the Credit Union's structural advantages have stayed exactly the same. Better rates. Member-first culture. Real relationships. All of it is still there. All of it depends on the member getting far enough into the process to feel it.

A 14-page application kills that before it starts.

A branded, mobile-first member portal live in five minutes. No LOS migration. That is where we start at Fuse.

Go to https://t.co/EWTkqzVVM1 and close some of that gap this week.

[Marc Escapa @escmarc]

I talk to Credit Unions all the time who went live on a new platform and never heard from the vendor again.

The implementation went fine. Go-live happened. And then nothing.

A year later, the team is running the platform at maybe 30% of what it's capable of. There are automations sitting idle, steps the AI could be handling, and nobody is showing them where to look.

The vendor's incentive ended at go-live. The Credit Union's real automation journey was just beginning.

At Fuse, go-live is the starting line.

Every two weeks, we show up and ask one question: what's the next thing we should tackle? Then we go build it.

That's how a Credit Union we recently worked with went from 62% automation to above 90%. The system kept improving because the partner kept showing up.

When did your current vendor last reach out with an idea you hadn't asked for?

If you're drawing a blank, send me a DM. I want to understand where your automation stalled or visit https://t.co/BAbaeotPI5

[Marc Escapa @escmarc]

Most Credit Unions don't realize how much choice they've given up until they try to add a vendor they actually like.

The legacy answer: use the one we've already integrated. Take it or leave it.

That's Vendor Lock-In, and it's not the Credit Union's fault. It's a business model that turns your own relationships into someone else's leverage.

Matthew Utesch at Canopy Credit Union wanted specific partners, including ones we'd never integrated. We added them to the build. No obscene fee.

The best vendor for the job should win. Not the one that got there first.

[Marc Escapa @escmarc]

The difference between a vendor and a business partner is often found in the development roadmap.

I recently sat down with the team at Navigant Credit Union to discuss why the "set it and forget it" approach to software fails in modern lending.

This is what we heard:

Many institutions are used to hearing "no" when they request specific adjustments to their lending workflow or data pages.

True partnership requires a collaborative rhythm where the Credit Union actually has a seat at the table with the development team.

When the technology is built to meet your specific functionality, it stops being a static tool and starts being a shared investment in your success.

If you feel like your current LOS roadmap is moving in a different direction than your business, DM me I’d be interested to hear how you are managing that gap.

https://t.co/BAbaeotPI5

[Marc Escapa @escmarc]

The oldest pitch in lending software goes like this: here's the box, here's what it does, you're done.

Need a change your own policy requires? Sorry. That's what's built.

I call it "the Black Box". And the reason it feels so frustrating isn't the Credit Union. It's that the product was finished before it ever met you.

Charlotte Nemec at Canopy Credit Union lived both versions. The old one handed her a fixed system and called rigidity normal. The new one bends to her process.

Software that can't flex to how a lender actually works was never a product. It was a ceiling we all mistook for normal.

[Marc Escapa @escmarc]

I have been inside a lot of Credit Unions this past year.

And the thing that stays with me is the people.

I watched a loan officer recognize a member walking through the door and pull up their file before they reached the desk. She already knew his situation.
She knew what he was probably coming in for. She pulled the file because she cared about having the right answer ready.

Fintechs lack that personalization and customer service in their App.

What I keep thinking about is how much of that is getting quietly lost.

The officer who would have had that conversation is now spending 40 minutes clicking between screens to get one file funded.

The warmth and the instinct are still there. The software is just burying it.

The cost of a broken process goes beyond cycle time. It is the conversation that never happened because the rep was buried in a queue.

[Marc Escapa @escmarc]

I talk to Credit Union lending leaders every week who launched a new tool six months ago and can't point to a single number that moved.

The vendor delivered the implementation. Invoiced. Left.

What I've found is that the real work starts after go-live. Most vendors don't price for that part and don't show up for it either.

At Fuse, staying after launch is the whole model. Because that's where Credit Unions actually need the help.

If your team launched something in the last year and the results aren't there, visit https://t.co/BAbaeotPI5 or DM me.

[Marc Escapa @escmarc]

A funder at Canopy Credit Union used to spend 30 minutes on every indirect loan. Multiply that across a busy month and you are burning weeks of someone's life on work that adds nothing.

Worse: the person doing it sat in finance. It was never even her job.
That's the Manual Funding Tax. Most Credit Unions pay it without ever seeing the invoice.

Here's what changed. In January, Canopy had its second busiest indirect month in three years. 92 loans. Funding now takes about 5 minutes each.

Matthew Utesch's words stuck with me: the record month didn't feel like anything.

That's what automation is supposed to feel like.

[Marc Escapa @escmarc]

I worked with a Credit Union that approved two new hires to handle growing application volume. Six months later I was back in the same room and the backlog conversation was identical.

Every person who joined learned the workflow from the team already running it. The workarounds got passed along. Volume kept climbing and the process didn't change to match it.

A Credit Union came to Fuse at 4% auto-processing had the same headcount when they reached 82%. The AI ran the tasks. The people ran the exceptions. The experienced members of the team finally had room to work the files that actually needed them.

Before the next headcount request goes to the board, ask which steps in the workflow could come out first.

Visit https://t.co/BAbaeotPI5 if you are dealing with this right now.

[Marc Escapa @escmarc]

Every Credit Union I talk to is quietly fighting the same battle: how do you grow without growing the team?

For decades, the answer from legacy software was a shrug. You scale, you hit a wall, and eventually the vendor tells you the quiet part out loud: you outgrew us.

That's the Headcount Trap. And it was never the Credit Union's fault. It's what happens when your system grows to a point and then becomes the reason you can't.

Charlotte Nemec, CEO of Canopy Credit Union, said it plainly. She doesn't want software that caps her. She wants one that grows with her.

So do we.