Hi fintech fam 💜

Last week, I pulled up to Unit21’s Risky Business event with a mic in hand, ready to moderate a panel of actual fraud-fighting legends from Brex and Gusto.

We weren’t there to sugarcoat AI or worship at the altar of automation. Nah. We got real.

Like—who’s actually using AI in FinCrime ops? Who’s just talking about it on LinkedIn? And what tech is a complete ✨delusion✨?

(Spoiler: if your “AI strategy” doesn’t start with clean data, Brex’s Flora Zhang has some words for you. And Gusto’s John Wiethorn? He’s out here gamifying compliance and comparing SARs to Minority Report. The man understood the assignment.)

And if you’re a Fintech Mavericks listener — or a founder building in the trenches — Brex built something just for you. 👉 brex.com/fintechmavericks

Custom perks, smarter spend, and a card that works as hard as you do. You’re welcome.

Now, let’s get into the real takeaways — AI, elbow grease, and the frameworks that matter when the stakes are high.

INNOVATION

AI, Elbow Grease, And The Future Of Fraud Fighting

Raise your hand if you've used AI at work this month.

Now keep it raised if you actually trust it to make a business decision without you double-checking it.

At Unit21’s Risky Business: A FinCrimeOps Tour event, nearly every hand in the audience went up when I asked the first prompt. 

Almost all dropped on the second.

That gap—between usage and trust—is where the real story of AI in financial crime operations begins. 

And it was the central theme of a panel I moderated last week at Unit21’s invite-only event gathering top fraud, compliance and fincrime leaders in one room. 

Joining me on stage were two heavy-hitters with deep operational chops: Flora Zhang, who leads Strategy, Operations, Risk, and Analytics at Brex, and John Wiethorn, Head of Financial Crime Compliance & AML/BSA Officer at Gusto.

Here are the big takeaways—and the tactical, spicy truths you won’t get at most events.

The Fraudsters Have Quarterly Planning Too

"AI is making the bad actors evolve faster than ever," Zhang said. “They’re using LLMs to social-engineer their way past our controls. Voice mimicry. Email spoofing. MFA isn't enough anymore.”

For Brex, the risk is amplified by the sheer volume of money flowing through a single B2B account. One compromised email can unlock millions.

Wiethorn agreed: “We have to stop underestimating how coordinated fraud is,” he said. “Bad actors do quarterly planning. They’re adapting to our defenses just as fast as we’re building them.”

This comment drew murmurs and head nods from the audience.

It forced a shift in perspective.

We often talk about fraudsters as if they were opportunists. But in 2025, they’re organized, resourced, and agile—more startup than street criminal.

And they’re not waiting for your new detection model to go live. They’re already testing it.

The implication? Security strategies need to evolve just as fast. That means moving away from rigid annual planning cycles and toward more dynamic, iterative models—borrowing a page from the product playbook. 

The companies that’ll survive aren’t the ones with the biggest budgets. 

They’re the ones who can ship updates the fastest.

Where AI Works (Hint: It’s Boring)

The room was filled with Chief Compliance and FinCrime Ops executives from fintechs and banks experimenting with AI across all corners of their risk stacks. But when asked where AI actually delivers value today, the answers were surprisingly simple.

"Enhanced due diligence," Wiethorn said. “Instead of analysts Googling names, parsing PDFs, and saving screenshots, AI now does the research, ranks results, and attaches relevant findings to the case file.”

Efficiency. Accuracy. Audit trail included.

This isn’t revolutionary—but it is scalable. For most risk teams drowning in alert volume, shaving even 10 minutes off per case can add up to hundreds of hours per quarter. That’s not just cost savings. It’s human energy reallocated to higher-stakes decisions.

Zhang highlighted narrow, well-scoped use cases: “Verifying an address. Checking if it’s virtual. That’s where AI shines.”

But ask it to write a Suspicious Activity Report (SAR) narrative? It collapses.

SAR narratives should clearly detail the "who, what, when, where, why, and how" of the suspicious activity.

“It doesn’t know what’s relevant,” Wiethorn said. “And relevance is everything.”

This is a critical distinction. AI is powerful at surfacing data. But the work of contextualizing it—making meaning from noise—is still very human. 

At least for now.

If You Want Good AI, Start With Good Data

The biggest blocker to AI? It’s not regulation. It’s internal hygiene.

"Everyone wants to use AI," Zhang said, “but no one wants to clean their data.”

That line got laughs. But also winces.

Brex, like many fintechs, has accumulated a Frankenstein stack of device intelligence over the years. Before deploying AI, Zhang’s team had to consolidate, reconcile, and backfill multiple overlapping signals.

It wasn’t sexy work—but it was necessary.

“Garbage in, garbage out,” she said.

This isn’t just a technical problem. It’s cultural. Many risk teams lack dedicated data resources. Analysts are forced to moonlight as engineers, building brittle logic atop patchy infrastructure.

If AI is going to be more than a toy—or worse, a liability—companies need to invest in foundational data work.

Think: clear taxonomies, unified identifiers, and a shared understanding of what a ‘high-risk transaction’ even means across teams.

You don’t need perfect data. But you do need accountable data.

The Frameworks That Actually Work

AI isn’t a silver bullet. It’s a tool—and like any tool, it needs a framework.

Zhang’s model for FinCrime ops decision-making is refreshingly simple:

1. Deterministic decision? Automate it.

2. Pattern-based? Consider AI—with human review.

3. Judgment-based? Keep it human.

Why does this matter? Because it gives teams a language to talk about AI—not just as tech, but as part of a process. It helps demystify decisions and gives non-technical stakeholders a way to weigh in.

Wiethorn’s rule of thumb was equally telling: “The moment I say I need headcount, leadership asks, ‘Can we build an agent instead?’ So I’ve learned to reverse-engineer from that question.”

In other words: If you want to scale your team, be ready to justify every human task. 

And make sure AI is solving a clear problem—not just tacked on because it's trendy.

The Most Useful (And Wildest) Ideas From The Panel

If you're building for 2030, here’s what should be on your radar:

• AI with memory. Zhang wants a model that can recognize patterns over time, not just isolated anomalies. “Something that can reason across multiple behaviors—and remember them.”

• AI-powered customer comms. Wiethorn believes support emails, RFIs, and KYB workflows will be AI-authored and human-sounding within five years. “We’ll cut out 80% of human touchpoints without anyone noticing.”

• Gamified compliance tools. His sleeper hit: Duolingo-style dashboards for analysts. Think badges for QA accuracy, streaks for SAR filings, leaderboards. “Why shouldn’t compliance be fun?”

• Never let AI close accounts. Both panelists were adamant. “Do it, and you’re one New York Times story away from regulatory hell,” Wiethorn warned.

This last point is worth expanding on. 

When AI makes customer-impacting decisions, the cost of error isn’t just operational—it’s reputational. 

One false positive can lead to media blowback, lost trust, and even regulatory scrutiny. 

The panel's advice? Use AI to assist. Never to judge.

Why It Matters

There’s no doubt AI is transforming financial crime operations. But that transformation is slower, more tactical, and more nuanced than the hype cycle suggests.

The best teams in the space aren’t “AI-first.” They’re problem-first. And they’re building:

• Systems that augment, not replace, human insight

• Controls with guardrails, not blind trust

• Playbooks that evolve as the threat landscape shifts

The fraud fighters who will win the next decade are already doing this work. They’re not chasing trends. They’re making thoughtful bets. They’re building taste.

They know when to trust automation—and when to roll up their sleeves.

Fintech Is Femme Leadership: AI Edition

Why AI, Why Now

To change the story of AI in finance, we need to change the storytellers.

This year, the Fintech Is Femme Leadership Summit goes AI Edition — not chasing the hype, but unlocking the real strategies to scale, raise, and lead with AI in 2025.

Think tactical frameworks, unfiltered conversations, and founders showing exactly how they’re weaving AI into product, GTM, security, and growth.

Who’s in the Room?

✨ Fintech founders scaling, fundraising, and building for the future

✨ Investors sourcing the next breakout AI-native fintechs

✨ Operators + execs where finance meets technology (and disruption)

What You’ll Get

🚀 Transparent founder stories

🚀 Investor insights you actually need

🚀 Playbooks that work (and what doesn’t)

🚀 Curated networking that puts you in front of the people who can change your company’s trajectory

👉 Only 200 seats. 

This isn’t just another AI panel. This is where fintech’s boldest leaders rewrite the playbook.

See you in San Francisco.

_{I WANT IT, I GOT IT}

• 🎧 Today’s Listen: Last week on Fintech Mavericks, I sat down with Andrew Gazdecki, Founder & CEO of Acquire.com, who has helped thousands of startups get acquired and facilitated over $500 million in closed deals. From building companies that no one else would, to redefining how acquisitions get done, Andrew has become the go-to voice for founders looking to exit on their own terms. Tune in here or wherever you get your podcasts.

• 🚀 Today’s Growth Hack: The Artist’s Way by Julia Cameron has officially made its way onto my reading list — and wow, I get the hype. I’ve just started working through the workbook, and I’m already hooked on the practice of Morning Pages. There’s something so freeing about waking up, dumping every raw, messy, unfiltered thought onto the page, and just letting it be. No editing. No overthinking. Just space to meet yourself before the world chimes in.

• 🧘‍♀️This week’s reminder: You don’t have to build alone. I’m giving major love to The Academy of Fintech — our private community for the founders, operators, and investors shaping what’s next in finance and prioritizing their sanity while doing it. Whether you’re navigating growth, chasing product-market fit, hiring your first (or fiftieth) team member — this is where the sharpest minds in fintech come together. Think: peer support, behind-the-scenes playbooks, and live sessions with the folks actually doing the work. Sound like your crew? Apply to join us here — we’d love to meet you.

_FINTUNES

🎵 Maisie Peters — You’re Just a Boy (And I’m Kinda the Man)

_{LET’S CONNECT}

📰 Share this newsletter with a friend and start growing your network.

🔗 Connect with me on LinkedIn for daily insights on female leadership.

🤝 Grow your business through content & community by partnering with me.

📣 Promote yourself to 50,000 subscribers by sponsoring this newsletter.

🎤 Host an epic event by booking me as a speaker, moderator, or emcee.

📚 Increase your expertise by ordering your copy of my book, Fintech Feminists: Increasing Inclusion, Redefining Innovation, and Changing the Future for Women Around the World.

That’s all for now! See you Thursday!

Love,

Nicole 💜