Quick summary:
- AI agents now read and interpret financial regulations across all 50 states in real-time, a task that previously required $250,000 legal surveys that were outdated the moment they were delivered
- Gen Z consumers are making nearly half their purchases using AI tools, creating pressure on merchants to provide real-time data that AI systems can access
- Stablecoin payment adoption surged 350% year over year as merchants discovered they could accept payments from 120 countries without managing local payment methods
Financial services companies face a practical problem. AI promises to automate compliance, payments, fraud detection, and more, but most implementations fail because the technology can’t handle the complexity of real banking operations.
At Money 20/20 in Las Vegas, The Disruption Is Now host Greg Matusky sat down with six AI leaders who built systems that actually work
The guests include:
- Chris Hilliard, Co-founder and CEO, Winnow
- Adam Gray, Chief Transformation Officer, Stax
- Daniel Aguilar, Head of Business Development and Alliances U.S, Prometeo
- Casey Williams, Fintech Investor and Operator, Fenway Summer
- Eleni Steinman, Co-founder and CEO, Loop Crypto
- Todd Rovak, Co-founder and CEO, Carefull
The conversations reveal how AI agents handle everything from reading dense legal documents to opening bank accounts across borders to catching elder fraud before it happens.
Watch now:
Key takeaways:
Legal AI only works when humans write the content it searches
Most AI systems fail at legal work because they search the open internet, where accuracy matters less than ranking high in search results. Winnow solves this by having attorneys build a database — they’ve spent over 64,000 hours so far — that covers every financial services regulation except student loans.
The platform works across departments that previously couldn’t share information. General counsels use it differently than compliance officers, who use it differently than enterprise risk teams. All of them get answers specific to their license type, charter, and product offerings.
Gen Z buying behavior forces merchants to feed AI systems
Gray saw a statistic that changed how Stax handles merchant services. Nearly 50% of Gen Z consumers now use generative AI tools when making purchases. Two years ago, that number was essentially zero.
The shift creates immediate pressure on payment processors. Merchants need their transaction data structured so AI systems can access it easily. When a consumer asks ChatGPT to compare prices or find the best deal, the merchant whose data feeds that answer wins the sale.
Stax responded by making data more accessible while avoiding the trap of monetizing it through proprietary services. The approach lets merchants partner with specialized marketing and analytics companies without being locked into their payment processor’s limited toolkit. As AI buying behavior accelerates, the merchants who share clean data capture the customers.
AI agents open bank accounts and fix payment errors humans create
Aguilar watched the same pattern repeat across Latin America. A business in the U.S. sends money to a vendor in Colombia. The bank doesn’t recognize the receiving institution. The payment gets rejected. The sender pays fees anyway, then must fix the error and pay fees again.
Prometeo built AI agents that communicate with other AI agents to solve the problem. The agents can open accounts, verify account details, transfer money, and handle compliance checks across 11 Latin American countries, each with different ID systems and verification requirements.
The agents work through APIs, creating infrastructure that handles what humans consistently mess up. In Chile, there’s one type of ID. In Colombia, there are seven or eight different kinds. The AI handles name matching, account verification, and compliance across all variations.
Wealth advisors won’t disappear but the ones who adopt AI will dominate
Williams invests in wealth tech startups through Fenway Summer. She sees AI transforming advisor workflows while making the human relationship more important.
Back-office automation handles routine tasks. AI generates first drafts of financial plans. Advisors spend their time on the emotionally complex work that actually drives the business: legacy planning, family dynamics, estate discussions.
The firms that win will be those that use AI to expand their client base downmarket while providing deeper services to existing clients.
Stablecoins eliminate the cross-border payment rejection problem
Steinman built Loop Crypto after watching merchants struggle to accept payments from international customers. The friction showed up everywhere. Someone in Vietnam wants to buy a subscription service from a U.S. company but doesn’t have a credit card. A customer in Poland tries to pay but the merchant’s payment processor doesn’t support that country’s local methods.
Her company processes crypto and stablecoin payments that eliminate these barriers. Transactions grew 350% year over year, with 85% of volume coming from outside the U.S.
The breakthrough is recurring payments. Most crypto options require customers to click a button for each transaction, creating friction that kills subscription businesses. Loop Crypto built smart contracts that let customers authorize automatic withdrawals, just like traditional autopay.
Banks are starting to notice. Steinman predicts that within a year, traditional fintechs will be moving money on stablecoin rails behind the scenes, even if consumers never know it.
Elder fraud AI catches behavioral patterns before money moves
Rovak built Carefull after realizing that two-thirds of bank deposits belong to older adults, yet every fintech product targets younger generations. The gap represented both a commercial opportunity and a human need.
Elder fraud costs $61 billion annually. However, only one in 44 incidents gets reported. Two-thirds of cases involve someone the victim trusts.
Carefull’s Gray Matter AI monitors transaction patterns and relationship changes that signal trouble weeks before suspicious payments appear. A new romantic relationship combined with unusual late-night account access might indicate a romance scam. First-time cryptocurrency purchases to unknown recipients could mean an investment fraud is developing.
“Gray Matter AI can understand what normal looks like for that person because it was based on real deviation,” Rovak explains. “It looks for behaviors that the banks and wealth management companies don’t look for, but it provides that understanding because what happens with elder fraud and senior exploitation is it happens way before they want to do a transaction.”
The platform gives adult children visibility into their parents’ financial activity without violating privacy or removing agency. Wealth managers and banks offer it as a value-add that keeps families connected and assets from leaking to competitors when the next generation gets involved.
Key moments:
- How Winnow’s legal AI uses only attorney-reviewed content to eliminate hallucinations (2:25)
- Why merchants need to change their strategy for Gen Z consumers (13:29)
- The cross-border payment rejection problem costing businesses double fees (17:14)
- The wealth management succession crisis hitting 60-year-old advisors (24:34)
- Using AI to keep below-minimum prospects warm until they become ideal clients (25:23)
- Stablecoin transaction volume surging 350% year over year with 85% from outside U.S. (30:12)
- How behavioral AI spots romance scams weeks before money moves (38:41)
- Why $61 billion in elder fraud happens before suspicious transactions appear (39:08)
Q&A
Q: How does Winnow ensure accuracy when most legal AI systems hallucinate?
Hilliard: We founded the company with the Buckley law firm, and we had access to over 140 attorneys. We’ve spent over 64,000 hours as of this morning building Winnow for just the information.
That’s all attorney content. You don’t have to worry about that when you go in and use AI. When you’re asking plain language questions and getting plain language answers, that’s just our data set. So, it’s just a walled garden of attorney researched, reviewed, and published content.
Q: How many in Gen Z use AI for purchases?
Gray: Just under 50% of Gen Z use generative AI to make purchases. Two years ago, they didn’t even know what that means. That’s insane. That half of that generation is now using these tools to make buying decisions. We need to provide the right information to software companies so they can fuel AI models with clean data.
Q: Will AI replace wealth advisors or enhance them?
Williams: I completely believe that advisors are not going to be disrupted but enhanced, and the ones who really adopt it and use it well will win. Because at the end of the day, people do still want to have a conversation with a human that they trust. And that level of trust and relationship is very hard to replicate with your phone.
Q: How are stablecoins used for cross-border payments?
Steinman: It’s actually quite difficult for a merchant to accept a local payment method. Most merchants in the U.S. say pay me with a credit card, but if you’re living in Vietnam, you might not have a credit card, but you do have crypto.
And a merchant can add stablecoin or crypto acceptance and now be seen as a local payment method in about 120 countries.
Q: How does Carefull catch elder fraud before suspicious transactions happen?
Rovak: Gray Matter AI can understand what normal looks like for that person based on behavioral deviation. What happens with elder fraud and senior exploitation is it happens way before they want to do a transaction. It’s a new relationship. It’s a first crypto payment to someone they may or may not know. It’s an investment scam. All this stuff happens and can be seen through earlier signals around their behavior.
