The Birth of AI First Fintech

Hey Fintech Nerds 👋

Block is offering consumer loans via its bank, Moonpay acquired Iron for $100m, and the FTC reported consumer scams grew 125% YoY. (Things to Know 👀). And just as I hit send, Klarna filled its IPO.

Those stories are such a good microcosm of the market.

Mobile-first-2010s Fintech is growing up, consolidating market share, and looking to own more economics. Stablecoins are the next big payments rail, and we’re in a scamdemic.

But what comes next in consumer Fintech?

AI Native Fintech.

And it’s going to be more transformational than the mobile-first era of the 2010s

That’s your Rant this week.

Here's this week's Brainfood in summary

📣 Rant: The Birth of AI First Fintech

💸 4 Fintech Companies:

1. Payman - The Agentic Payments API

2. Mansa Finance - Liquidity for Cross border payments

3. Stablecore - The "side core" for stablecoins

4. MyTu - The AI-first digital neobank (EU)

👀 Things to Know:

1. Block subsidiary approved to offer consumer loans by FDIC

2. Consumers lost $12.6bn to scams last year, up 25%

3. MoonPay acquires Iron to add enterprise stablecoin solutions

📚 Good Read: Scam Empire - The investigation into $275m worth of investment advisor scams

Weekly Rant 📣

The Birth of AI First Fintech

Imagine saying, "File my taxes," and it just happens. Or getting a message: "Six invoices are overdue today. Would you like me to chase them?" No forms, no apps to navigate, no specialized knowledge required. Just you, speaking normally, with an AI that understands your financial life intimately.

This is the coming era of AI-native finance, where vibe is the UI and productivity per employee is closer to $10m than $100,000.

What happens when every consumer can create their own financial experience just as we do with modern coding tools? We get liquid software. It adapts to users. True personalization. Unlimited experimentation.

Thesis: We have passed peak AI sentiment, but we're just getting started on unlocking AI value. I believe AI Native Finance will be bigger than the mobile/cloud era – transforming both how financial products are built and how we interact with them.

The Fintech Opportunity has never been greater

Fintech is just 3% of the total market cap of financial services.

If AI does one thing, it fundamentally alters the cost structure of building and running a company.

It also lets these tiny efficient companies grow incredibly fast.

AI has allowed companies to reach $10m and $100m ARR with teams of 10 people faster than ever, speeding up the cycle of new company formation. Incumbents with 10,000+ teams and billions in revenue will find hundreds of competitors popping up.

This will be massively disruptive to financial services over the long term.

People overstate the pace of change and understate the impact.

AI is very good at the sort of problems finance is plagued by

So much of financial services live in unstructured, messy formats like

• PDFs for contracts, identity documents, compliance policies

• Flat files for payment or credit messages (e.g. ISO8583)

• Emails and spreadsheets for everything, and the Sharepoint, will someone please remember the Sharepoint!

These unstructured documents persist because, contrary to appearances, financial services is not a calculator. It requires a significant amount of judgment based on the unique circumstances of a customer, the market, and the product being offered.

Automating the messiness of judgment and the ability to handle edge cases is almost impossible.

Whenever I write about a new AI thing, the most common feedback is "why can't RPA or RegEx do this?" This is a sure-fire sign that the person hasn't tried to write code to handle the sheer volume of edge cases in things like invoices or spreadsheets. Every spreadsheet is idiosyncratic. Is the price $10 or is the unit 10? In what language? And where on the file system is that spreadsheet stored?  

The way you make software work is by constraining the variables.

Robotic Process Automation works best when there's a process.

What's the process for the edge case where the PDF wasn't attached to the email? Or they sent a spreadsheet instead of the PDF? Or they sent it zipped?

You could handle each of these individually, but it takes time and energy.

LLMs and AI Agents are generally very good at this sort of messy edge case management. In turn, that means you can build software much faster and much cheaper than ever.

AI Native Fintech alters the economics of the industry.

Whenever you alter the cost structure of an industry, it benefits incumbents but not nearly as much as the disruptors. Just look at the cost per account of an incumbent bank to operate a checking account (~$250 per year) vs, say Nubank (~$10 per year).

There's a chasm between new technology natives and new technology adopters:

• Mobile Banking (Incumbents) vs Mobile Native Banking (Chime, Nubank, Revolut)

• AI Banking (Incumbents) vs AI Native Banking (“Do my taxes,” “chase my invoices”)

Where this gets exciting is when you look at the revenue per employee and the productivity revolution possible.

From AI of the Coast

It continues

But cost efficiency is just the foundation. The truly revolutionary aspect isn't what happens behind the scenes—it's what happens when AI finally fixes the broken interface between humans and their financial lives.

The Vibe UI: Talk to Me, Don't Make Me Talk to You

What is the AI native customer experience? 

It is almost certainly not the current chatbots. The trend gaining traction is the idea of “vibing” with your AI Agent. What is the Vibe UI?

Think about it in terms of the UI evolution from:

• The command line: Learn specific words and codes to use

• The graphical interface: Learn a mouse and keyboard combination

• The vibe user interface: Natural and doesn’t require a new skill

Traditional UIs demand you learn their language. Command lines need specific syntax. Apps need you to navigate their logic. Even voice assistants require "Hey Siri" or specific commands to work.

But Vibe UI flips this entirely. It learns your language, your context, your needs.

Imagine financial interfaces that actually get you:

• Show me, don't tell me: Instead of explaining your spending patterns in text, it visualizes them instantly. "How's my spending this month?" gets you a heat map of where your money went, with problem areas highlighted.

• Talk to me, don't make me talk to you: Rather than waiting for you to discover a late payment fee, it reaches out: "Hey, noticed your water bill is due tomorrow, but you're low on cash. Want me to move some money from savings to cover it?" It's anticipatory, not reactive.

• Feel me, don't process me: The difference between "Transaction declined: Insufficient funds" and "This purchase would put you $40 over. Your paycheck hits tomorrow—want to proceed anyway?" One is a process result; the other understands the human moment.

Alexa and Siri ruined our expectation of voice interfaces. And “show me” in graphs sounds like PFM, but there’s something powerful here about an AI agent that is able to give you something contextual, and unique to you. Don’t like graphs? Here’s a table. Don’t like tables? Here’s a two way conversation.

As Monzo Founder Tom Blomfeld tweeted recently

What makes the Vibe UI powerful for finance:

1. Contextual awareness: How the UI might anticipate needs based on location, time, spending patterns ("Your electric bill is higher than usual this month - should I analyze why?")

2. Emotional intelligence: How the UI might adapt its tone and approach based on the user's financial stress levels

3. Progressive disclosure: How complex financial information could be layered based on user interest ("That's a good question about your mortgage rate - would you like the simple answer or the detailed analysis?")

Vibe UI could create financial interfaces that blend seamlessly into your life—understanding context, predicting needs, and communicating like a financially savvy friend rather than a banking terminal.

Why memorize commands when your finance app should just get your vibe?

So when this becomes customer facing I imagine being able to speak things into existence. Simple “wishes” like 

• Payments: “Buy me the same groceries as last week and deliver when I’m home”

• Financial guidance: “Optimize my savings and investments”

• Taxes: “File my taxes” (can you imagine!!)

• Consumer cashflow: Agent says “hey, you’re running low on cash, would you like to move some money from savings so you don’t get a fee?”

• Business collections: “You have 6 invoices late today, would you like me to give them a call?

• Embedded Financial Planning: "Create a vacation fund and find me the optimal savings strategy based on my spending habits"

• Financial Identity Management: "Keep my financial identity current across all my accounts automatically"

• Personalized Financial Coaching: "Analyze my spending patterns and coach me on better habits" (goes beyond alerts to actual behavioral change)

All of this is possible. 

What’s unclear is what the user experience will be, but the experimentation will come from the AI Natives. 

But there’s a problem. Financial services is a regulated industry. Wherever you move there is regulation.

That’s why doing AI in Fintech is hard. Because law is hard. Compliance is hard.

AI is hard because of regulation.

I was at several dinners this past week at Fintech Meetup where the same themes came up from bankers around AI.

• Will the regulator allow it?

• How can we ever use a black box?

• What if it hallucinates?

These are valid questions, and the scrutiny on banks is enormous (as I've written before). 

But they're also blockers. When you have a large organization, questions like these cause paralysis. Not because people aren't smart but because of a series of other hidden constraints. Like

• Having an existing business, product and company to manage

• Having a calendar filled with meetings and committees

• Not being able to use cutting-edge models by security

• As a result of the above, not attracting builders who just do things

Meanwhile an AI native or Fintech company doesn’t exist unless it is doing something an incumbent would struggle to. Their margin is your opportunity.

But what if AI helps us be better at compliance?

Here’s a mindf*ck for you. 

Most laws, regulations, and compliance policies are like a giant prompt for an LLM.

Put another way, AI is a really great way to understand the giant documents that most humans would never read and fully understand. The reason we develop procedures and train staff in compliance is that we can’t expect them to read thousands of pages of policy documents and law.

We can expect that from an AI.

In fact, an AI performs much much better if you do give it that.

AI natives can feed the regulations directly to the AI, creating direct, specific policies and procedures tailored to their unique business. And this will all happen at machine speed. 

The irony? The very regulations designed to protect consumers might accelerate the AI native’s time to market, and lower the overall risk to consumers. 

Incumbents see compliance as a cost center. 

AI natives could see it as a competitive advantage.

The competitive dynamics will be fascinating.

• The incumbents have to move slower because they have a higher level of scrutiny, a lower risk appetite, and a lot of organizational process to overcome. 

• AI Native companies have much lower cost, and very little infrastructure or people to manage. They’ll move much faster. 

• The Fintech companies in the middle are the most interesting. If they can really take advantage and already have a customer base, they could be in a sweet spot but could also be threatened.

Speed is a competitive advantage. 

However as we learned from banking as a service, going fast is not always a good thing. Going too fast can mean:

• KYC becomes a conversion problem, and we end up with a ton of fraud and AML risk. 

• Due diligence gets missed, customers get harmed, bad credit gets issued.

The fact that AI can help you build very fast, doesn’t remove the scale of risk in the financial services industry. And if you’re a Fintech company who leans too aggressively into AI, explaining that to your partner bank can be incredibly challenging (as many are currently experiencing).

Making AI Native Fintech.

Mobile native fintech gave us zero-fee checking, credit building, and expense management that doesn't suck. 

But that was just removing friction from existing products. 

AI native fintech rethinks what the experience of finance is.

What if instead of developers building apps for users, users build their own experience? That world is one Alex Komoroske calls “liquid software.” Software we can craft, change and adapt in real time, with just words.

Every day I see demos of people creating a Spotify clone in seconds on apps like Bolt or Loveable.

Who’s going to build the consumer Fintech that lets them create their own software while managing compliance?

Ten years from now, we'll look back and realize that "filing taxes" or "managing investments" weren't actually things humans needed to do—they were just artifacts of poorly designed systems waiting for AI to fix them. 

The companies that understand this transformation will build products we can't yet imagine, but soon won't be able to live without. 

The gold rush is just beginning, Fintech Nerd.

ST.

PS. If you can create a Fintech company for 10x less than during the Banking as a Service boom, what happens if founders can bootstrap the next Chime? What happens when VC exits, and growth targets aren’t the incentive model anymore? Maybe that will be a net positive.

4 Fintech Companies

1. Payman - The Agentic Payments API

Payman is an API and SDK for agentic payments. It creates wallets for agents, and a sandbox for them to operate within. Everything in this sandbox is audit d, monitored, and can be flagged for human review. It has secured investment from Visa, Coinbase and Circle to build the most secure way to pay via AI.

🧠 What's the moat here? From their very obviously AI generated FAQs, it sounds like they're using open banking to pull money in and out of a stablecoin wallet. Then they use the stablecoin wallet + multisig transactions to ensure they have an audit trail. It's a smart model and an obvious target state (I've written about it before). But my question is  why will this team execute and sell? I reckon with a few smart payments nerds; we could get this built quickly and nail the go-to-market 🤷‍♂️

2. Mansa Finance - Liquidiy for Cross border payments

Mansa is a liquidity partner for companies that do cross-border trade. It uses Stablecoins to connect currency pairs and manage its FX ri ks. Mansa is growing into a Wise-like offering for businesses with multi-currency accounts and virtual cards.

🧠 The financing opportunity for emerging markets is incredible if you have the right local knowledge. The founder mentions that the collateral requirements are insane, creating a huge credit gap in the region. DeFi rails help bring liquidity to market, but where it often fails is the local writing skill. Mansa aims to help here. It feels like a founder-market fit if I ever saw it.

3. Stablecore - The "side core" for stablecoins

Stablecore allows banks to offer digital assets and issue stablecoins through their "side ore" platform. Financial institutions can also offer stablecoins as a payment method to complement ACH, RTP, and cards. It integrates digital asset custody and a blockchain infrastructure and links to the bank's existing core system of record (SOR).

🧠 The existing cores, like FIS, Jack Henry, and Fiser, should be scrambling to partner. When the GENIUS bill (or something like it) passes in 6 to 12 months, every bank will want a stablecoin solution. Piece together custody, brokerage, key management, and compliance is nontrivial. Companies like BVNK or Zerohash can help, but something focussed with "core" messaging is neat.

4. MyTu  - he AI-first digital neobank (EU)

MyTu offers a mobile-first digital bank account with no deposits or payment fees. It offers kids' and family accounts in a single app and has over 50,000 users signed up to date. MyTu aims to secure a full banking licence in Lithuania in the coming 12 months.

🧠 They've been around since 2019 and have 50k users? They're competing with some big names. There's a long road ahead. That said, more of these apps should be baked into kids' accounts as a default.

Things to know 👀

1. Block subsidiary approved to offer consumer loans by FDIC

Block’s industrial bank Square Financial Services, Inc. (SFS) has received approval to begin offering the consumer loan product Cash App Borrow from the FDIC. SFS is an independently governed subsidiary of Block Inc. Cash App Borrow loans are typically less than $100 and cost a fraction of the fee of an overdraft or payday loan. 

🧠 Block’s share price is down 33% YoY. 

• It’s where it was 3 years ago despite having 3x’d revenue.

• The great multiple reset is in progress.

• Lets see if “founder mode” Jack Dorsey can help bring some vigor and growth back or they’re on the road to being a value stock

🧠 This is not a new product, but now they get to own more of the economics.

• The product was historically offered via an external bank partner.

• Being a part of the group means Block keeps more of the economics

🧠 So I guess the bank holding company act doesn’t apply? 

• There’s a bright line between banks and tech companies historically.

• The compliance overhead of being a bank is massive

• In ZIRP era tech company multiples were much higher than bank stocks.

• Maybe that’s changing?

• The smaller ILC that serves the tech company is a model Apple could potentially replicate too. Keep your eye on Utah.

🧠 The path for big tech to become small banks?

• Industrial Loan Charters (ILCs) were designed for car companies and industry to have lending operations to businesses (like car dealerships).

• That’s getting stretched a bit here, but makes a lot of sense.

• As big tech does more finance, they’re not quite full fat banks, but this is a nice middle ground.

2. Consumers lost $12.6bn to scams last year, up 2 %

The FTC says consumers lost $12.5bn to scams last year, driven by social media, AI, and Crypto investment schemes. That's a massive jump from 2023, when just $2.5bn of scams were reported. Imposter scams are the most common form, and bad actors tend to be romantic partners, family members in distress, or tech support experts.

🧠 That's a 125% CAGR since 2023 (!!).

• These scam typologies have been around forever, but never at this scale.

• AI has made them cheap and easy to reproduce.

• Social media makes consumers easy to access,

• and faster payments means faster fraud

• RTP also means there's often no clawback mechanism banks can use.

• That's why I often say we're in a scamdemic™ - and I'm passionate about / dedicating effort to trying to fix this mess in my $day_job

🧠 UX best practices exist.

• In the UK several best practices have emerged.

• If Revolut suspects a scam, it will require a customer to input more data, watch a scam awareness video, call Revolut, and even wait 24 hours before a payment can be sent.

• After all of that you have to be pretty sure you want to send a payment

🧠 Banks don't want liability, but they have to do something.

• The new administration might not force banks to take liability, but the courts and states could

• Customer losses and complaints will clog up the bank support lines

• Investing in scam detection is critical

• Pro tip, users' devices and their behavior on it is such a useful sign a scam might be about to happen.

• E.g., in tech support scams, you might see a lot of screenshots or remote screen-sharing software.

🧠 We need cross-sector data sharing and to play well together.

• Historically fraud data services like EWS were by the banks for the banks. No others allowed.

• But this issue is too big to remain that way.

• We need big tech, merchants, Fintech, and banks at the table, along with initiatives like Sardine* Sonar and Plaid Beacon.

• If we care about this industry we have to fix this mess.

• This isn't the time for pointy elbows and

3. MoonPay acquires Iron for $100m to add enterprise stablecoin solutions

Crypto on ramp provider Moonpay is acquiring API stablecoin infrastructure provider Iron. This will allow MoonPay to offer instant, global settlement to enterprises in merchants. Moonpay is positioned to compete with Stripe, BVNK and Zerohash who offer similar services and have seen significant growth in the past 12 months.

🧠 Stablecoins had a huge week. 

• Stripe CEO Patrick Collison testified to congress about the benefits of stablecoins, and

• The OCC gave clarity to digital banks to be able to offer crypto asset custody.

• At Fintech Meetup, I had 3 senior bankers ask me for help with their stablecoin strategy. Lots happening.

🧠Moonpay was the canonical onramp, can they compete in stablecoins? 

• Anyone who bought crypto in the NFT craze may have seen Moonpay.

• They were the default way to sell fiat and buy crypto. But the market has moved on.

🧠 The death of the onramp the beginning of stablecoins as the next big payments rail

• Who needs an onramp if speculative assets are crashing?

• It’s clear stablecoins have product market fit and are growth.

• The “stablecoin sandwich” is making the stablecoin and on / off ramp invisible to users

• Moonpay needs a play here to stay relevant (especially if they’re due a fund raise, M&A is a good story)

🧠 This is a great result for Iron. 

• The company product has been live with stablecoins for ~6 months after a pivot from building an on-chain bank.

🧠 What a difference a year makes. 

• We’re not far away from Money 2020 Europe in June.

• Last year I remember mentioning to lots of payments pro’s “keep an eye on stablecoins.”

• It seemed so far away from reality to many back then. Not anymore.

Good Reads 📚

1. Scam Empire - The investigation into $275m worth of investment advisor scams

Thirty media organizations uncovered a scam operation with 32,000 victims run by scam call centers. Agents convinced users to “invest” in financial products, but the catch is the victim can never withdraw their “profits.” One group in Cyprus employed more than 480 people alone.

🧠 At least 500,000 people are working in scam call centers in Cambodia. I spoke to the CEO of the Merchant Risk Council this week who shared this statistic. Many of these “scammers” have been human trafficked and are working under modern slavery conditions.

🧠 Investment advisor attacks are consistently effective. It looks like everything is going well for the victim until it doesn’t.

🧠 This is a top of mind issue for all regulators and compliance teams. I’ve spoken to three different agencies about the issue this week.

🧠 We’re in a scamdemic. I’m going to keep saying it. This is the biggest issue in our industry.

Tweets of the week 🕊

That's all, folks. 👋

If you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)

(1) All content and views expressed here are the author's personal opinions and do not reflect the views of any of their employers or employees.

(2) All companies or assets mentioned by the author in which the author has a personal and/or financial interest are denoted with a *. None of the above constitutes investment advice, and you should seek independent advice before making any investment decisions.

(3) Any companies mentioned are top of mind and used for illus rative purposes only.

(4) A team of researchers has not rigorously fact-checked this. Don't take it as gospel—strong opinions weakly held

(5) Citations may be missiI'veand I've done my best to cite, but I will always aim to update and correct the live version where possible. If I cited you and got the referencing wrong, please reach out