Daniel sent us this one. He was reflecting on our episode about the state of retail banking in Israel, and he landed on something that I think is actually a much bigger idea than it first appears. The core question is this: what if we stopped asking banks to be good at customer experience, something they are structurally terrible at, and instead decoupled the whole thing? Let the banks run the settlement plumbing, the interbank transfers, the ledger maintenance, and let a new layer of fintech intermediaries build the actual user experience on top. Daniel's asking whether that model is viable, whether it's being pursued anywhere else, and what kinds of companies would flood the market if a full open banking API were actually implemented somewhere like Israel.
The reason this question lands with such force is that we have a perfect case study in dysfunction sitting right in front of us. I want to start with something concrete, because it's almost too absurd to believe without the detail. Cal is one of Israel's main card-issuing companies. Their call center closes at four PM. Four in the afternoon. If you have a problem with your credit card at four oh five, you are talking to nobody until the next morning. This is not nineteen eighty-five. This is now. And that is the baseline we are working from.
It's the kind of detail that sounds like a bit, but it's not. I discovered it myself on the day of recording. I genuinely had to check the website twice because I assumed I was reading the Friday hours or something. That's the standard closing time. And Daniel's point is that this is not a glitch. It's a symptom. The entire system is built around the convenience of the institutions, not the customers, and no amount of regulatory nudging seems to change that.
To understand why that four PM closure is more than just an inconvenience, it's a structural symptom, we need to look at what happened with the Strum Committee and Israel's open banking efforts. The Strum Committee was convened in twenty sixteen and reported in twenty seventeen. It was a major government initiative to increase competition in Israeli banking. The headline finding was stark: five banking groups control roughly ninety-five percent of household credit in Israel. That's an oligopoly by any definition. The committee made a series of recommendations, and I want to be precise here because the details matter. They recommended structural separation of credit card companies from the banks, they pushed for new bank entrants, and they called for open banking APIs that would let third parties access customer financial data with consent.
What actually happened?
The credit card separation happened, sort of. Cal and Isracard were divested from their parent banks, but the open banking piece has been a slow-motion disappointment. The Bank of Israel did issue an open banking API regulation with staged implementation from twenty twenty-one through twenty twenty-three. On paper, Israel adopted open banking. In practice, the banks dragged their feet. The API endpoints that were implemented were limited, mostly read access to basic account information. The documentation was poor. There was no real third-party integration ecosystem that emerged. And critically, the write access endpoints, the ones that would let a fintech actually initiate payments or open accounts on behalf of a user, were not fully mandated. They were left vague.
The banks technically complied with the regulation without actually enabling the thing the regulation was designed to create.
They built the door but didn't hand over the keys. And this is not a uniquely Israeli problem. We've seen variations of this in the United States, in Canada, in Australia. Banks everywhere treat open banking as a compliance checkbox, not as infrastructure they actually want third parties to use. But Israel's concentrated market makes the problem particularly acute. When five groups control ninety-five percent of household credit, there's no competitive pressure to improve the consumer experience. You don't lose customers by having a call center that closes at four PM, because where are they going to go? To another bank with a call center that closes at four thirty?
The question becomes: if forcing banks to be better at UX hasn't worked, and it's been almost a decade of trying since Strum, what if we stopped asking them to do that at all? Let's look at the decoupling model. How it would work technically, what the pieces are, and why it hasn't happened yet.
Let's break this down. The banking function, what we think of as a bank, is actually two completely different things bundled together. On one side you have settlement infrastructure. This is the interbank transfer system, card network clearing, account ledger maintenance, the actual movement of money between databases. It is deeply unglamorous, heavily regulated, and hard to do well. On the other side you have the customer experience layer. The mobile app, the notifications, the budgeting tools, the customer support, the account opening flow, the interface where you actually see your money and tell it what to do. These two things are currently bundled inside every bank. When you open the Leumi app or the Hapoalim app, you are talking to both layers at once. The decoupling model says: split them.
The mechanism for splitting them is a full open banking API.
And I want to be specific about what full means here, because the phrase open banking gets thrown around and most people think it's about checking your balance in a third-party budgeting app. That's the shallow version. A full open banking API needs two categories of access. First, read access: transaction history, account balances, standing orders, direct debit mandates, loan details. This is what Israel's regulation partially delivered. Second, and this is the transformative piece, write access: the ability for a third party to initiate a payment from your account, to open a new account on your behalf, to set up or modify a direct debit, to apply for credit using your transaction history as underwriting data. If a fintech has both read and write access through standardized APIs, it can become what is effectively a front-end bank. It can offer the full banking experience without holding a banking license and without maintaining settlement infrastructure.
The fintech builds the app, handles the customer relationship, designs the interface, manages the support, and the underlying bank just processes the transactions and maintains the ledger.
The bank becomes a utility. The fintech becomes the face of banking. And this is not a hypothetical. Parts of this model exist in the UK and Europe under PSD2 and the Open Banking Implementation Entity, the OBIE. The UK created a three-tier model. Tier one: the Account Servicing Payment Service Providers, the ASPSPs, which is a terrible acronym for the banks. Tier two: Third Party Providers, the TPPs, which access data via APIs. Tier three: the customer, who can choose to use a TPP instead of their bank's own interface. But here's the crucial limitation. In practice, TPPs in the UK and EU are still heavily constrained. They can read data and they can initiate payments, but they can't build a full banking experience on top. They can't open accounts. They can't manage direct debits comprehensively. They're more like a window into your bank than a replacement for your bank's app.
Even the most advanced open banking regime in the world hasn't actually delivered the full decoupling Daniel is describing.
And the reason gets to the heart of why this hasn't happened in Israel either. Banks control the API gateways. They can set rate limits on how many API calls a fintech can make per minute. They can charge prohibitive fees per API call. They can simply not implement the write endpoints that would enable account opening. The Bank of Israel's regulation mandated read access but left write access vague. And even where write access technically exists, the banks can make it so expensive or so rate-limited that no viable business can be built on top of it.
The justification they always reach for is security. If we open up write access, fraud will explode.
The security counterargument is the go-to move for every bank resisting open banking. And let me be fair: there is a legitimate security consideration here. Broad API access does increase the surface area for potential fraud. You are creating more points of entry into the payment system. But the real risk is manageable, and the UK's experience demonstrates this. They've had mandated open banking APIs since twenty eighteen, with over five million users of open banking services as of twenty twenty-four. Their data shows that fraud rates on API-initiated payments are actually lower than on card-not-present transactions. Because API-based payments use strong customer authentication, SCA, which typically requires two-factor verification for each transaction. And because the access is tokenized, the fintech never sees your actual banking credentials. They get a token that grants specific, limited permissions.
The security argument is a smokescreen.
It's a legitimate concern that has been largely solved, which makes it a very convenient smokescreen. The real resistance is commercial. Banks do not want to lose the customer relationship. The customer relationship is where the data is, where the cross-selling happens, where the brand loyalty lives. If a fintech owns the interface, the bank becomes a commodity provider of settlement services. And while settlement is a real business, it's not the business that bank executives built their careers to run. Nobody goes into banking dreaming of being the invisible plumbing.
Yet, if you look at the actual economics, the plumbing business is not a bad business to be in. JPMorgan Chase's wholesale and commercial banking division generates higher profit margins than its consumer banking division. The invisible plumbing is actually more profitable per customer than the shiny consumer app.
That's a critical point. The margins in wholesale banking, treasury services, and settlement infrastructure are strong. And for many large banks, the consumer business is actually a cost center when you factor in branch networks, call centers, marketing, and regulatory compliance for millions of small accounts. If banks could shed the expensive consumer-facing operations and focus on the high-margin infrastructure business, they might actually be more profitable. But that requires a psychological shift that is very hard for incumbent institutions to make. They've been consumer brands for generations. Becoming a behind-the-scenes utility feels like a demotion, even if the numbers say otherwise.
Okay, so the model is technically feasible. The security objections are manageable. The commercial resistance is real but potentially shortsighted. But what would actually happen if it were implemented? What kinds of companies would emerge, and how would the incumbents respond?
Let's imagine the decoupling model is fully implemented somewhere like Israel. A genuine full open banking API with read and write access, reasonable pricing, standardized documentation, and regulatory backing. What categories of fintech would emerge? I think you'd see three distinct types. The first is the UX-first retail bank. This is a company that builds a complete banking experience, the mobile app, the notifications, the budgeting tools, the customer support, but uses the existing banks' settlement rails for everything on the back end. They compete entirely on interface, on speed, on support quality. Think of a Monzo or a Starling, but built on top of Leumi or Hapoalim's infrastructure rather than building their own from scratch. The barrier to entry drops dramatically because you don't need a banking license or a hundred million dollars in capital reserves. You just need a great product team and API access.
A call center that stays open past four PM, presumably.
That alone would be a competitive advantage in this market. The second category is specialized intermediaries for specific use cases. Payroll-linked lending, where a fintech sees your salary deposits and offers credit automatically based on that predictable income stream. Automated savings rules that sweep money into investment accounts based on your spending patterns. Subscription management tools that identify and cancel unused subscriptions across all your accounts. These are not full banks. They do one thing, but they do it with access to your complete financial picture, which makes them dramatically more effective than a standalone budgeting app that you have to manually update.
The third category?
B2B middleware that lets non-financial brands offer banking-like experiences without becoming banks. Imagine a major retailer, say a Shufersal or an IKEA, offering a branded current account inside their app. They're not becoming a bank. They're licensing the middleware, which connects to the settlement infrastructure, and they're putting their brand and their customer relationship on top. Or a telecom company, a Cellcom or a Partner, offering banking as part of their mobile plan. This is the most radical category because it means banking becomes a feature embedded in other products, not a standalone product you go to a bank to get.
The telecom analogy is actually the clearest way to think about this whole thing. Mobile virtual network operators, MVNOs like Giffgaff or Mint Mobile, they don't own any towers. They don't own spectrum licenses. They lease network access from the incumbents who built and maintain the physical infrastructure. And they compete entirely on pricing, on customer experience, on brand. The incumbents grumbled about it, but they took the wholesale revenue, and the MVNOs found their niche. The banking equivalent would be virtual retail banks that lease settlement access from the incumbents. The infrastructure owner still gets paid. The virtual operator competes on everything the customer actually sees and touches.
The parallel is remarkably clean. In telecom, the incumbents initially fought MVNOs, but eventually accepted them because the wholesale revenue was real and the alternative was regulatory intervention that might be worse. In banking, the same dynamic could play out. If banks refuse to cooperate voluntarily, the regulator eventually mandates access at regulated prices. The UK's Competition and Markets Authority did exactly this, mandating API pricing that was reasonable rather than prohibitive. The banks complained, but they complied, and the ecosystem started to grow.
There's a specific reason this model might work better in Israel than in many other places. The population is about nine and a half million. Smartphone penetration is eighty-eight percent. The banking sector is absurdly concentrated, five groups controlling roughly ninety-five percent of retail banking. That concentration is normally a problem, but in a decoupling scenario it becomes an opportunity. A single fintech with a great UX could capture significant market share very quickly if API access were real, because the addressable market is essentially everyone who is frustrated with their current bank, which is approximately everyone.
You wouldn't need to convince people to switch banks, which is famously one of the hardest things in consumer finance. People stay with bad banks because switching is painful. Account numbers change, direct debits need to be reestablished, there's a gap where you're not sure if your salary will land correctly. But if the fintech is building on top of the existing bank infrastructure, the underlying account might not even need to change. You're just changing the interface layer. That dramatically reduces switching friction.
Which brings us to the catch. Because there's always a catch. Banks would need to be compensated fairly for their infrastructure. They did build it, they do maintain it, and settlement is complex and expensive to operate. The problem is that the current fee structures, in the rare cases where banks even offer API access, are designed to be prohibitive. They're not priced to enable a market. They're priced to prevent one.
This is exactly what happened in the UK before the CMA stepped in. Banks were offering API access but charging per call at rates that made any consumer-facing business model impossible. If every time a user opens their app to check their balance it costs the fintech a few pence, and the user checks their balance ten times a day, the economics collapse before you even get to building features. A regulated fee schedule is essential. The UK's model essentially says: the marginal cost of an API call is near zero, so the pricing should reflect that. Banks can charge for value-added services, but basic data access and payment initiation need to be priced at or near cost.
Let me push on something. We've been talking about this as if the banks would grudgingly accept becoming invisible utilities. But is there a world where they actually embrace it? Where they look at the economics and say, you know what, running the plumbing at scale with predictable wholesale revenue is actually a better business than competing on consumer apps?
There are signs of this thinking emerging. JPMorgan has been building out its wholesale API business aggressively. They see that the future of their consumer business might be less about owning the customer relationship and more about providing the infrastructure that powers other people's customer relationships. In Israel, Bank Leumi launched Pepper, which is a digital bank. It's an interesting case study because it's an incumbent trying to build a good UX, but it's still owned by the bank. It's not a true third party. It's the bank competing with itself, sort of. The fact that they felt the need to create a separate brand suggests they know their main brand is not winning on experience.
Pepper is the exception that proves the rule. It exists because Leumi recognized that its own customer experience was insufficient for a certain segment of the market. But instead of fixing the main bank's experience, which would require changing everything about how the organization works, they spun off a separate thing. And even then, it's not open. It doesn't provide API access to third parties. It's a walled garden with a nicer app.
Compare this to Brazil, which has been more aggressive than Israel on open banking. Brazil implemented mandatory data sharing across all account types with standardized API specifications between twenty twenty-one and twenty twenty-four. It's a phased rollout that goes further than most countries, including mandated write access for payment initiation. Early results show fintechs like Nubank gaining share, and Nubank now has tens of millions of customers across Latin America. But the incumbents, Itaú and Bradesco, are still dominant. Open banking didn't destroy them. It just created space for new entrants alongside them.
The nightmare scenario the banks are protecting against, total disintermediation, doesn't actually happen even when the regulation is aggressive.
It doesn't. What happens is the market expands. More people access financial services in more ways. The incumbents lose some share but the total pie grows. And the incumbents who adapt, who build good API businesses and focus on their infrastructure strengths, they do fine. The ones who fight it lose relevance slowly.
We've got this vision of a decoupled banking system. But what does that mean for you, whether you're a customer, a developer, or someone thinking about building in this space?
The practical takeaway is this: the decoupling model is technically viable and has precedent in telecom and in parts of European open banking, but it requires regulatory force to overcome incumbent resistance. Voluntary cooperation has failed everywhere it's been tried. If you're in a country with stalled open banking, Israel, the United States, Canada, the key metric to watch is not whether APIs exist. It's whether write access, account opening and payment initiation, is mandated and priced reasonably. Read access without write access is a window. Write access is a door. And banks will build windows all day long if it means they don't have to build doors.
If you're a developer or a fintech founder, the technical standards already exist. The Bank of Israel has an API sandbox. The UK's OBIE specifications are publicly available and well documented. The bottleneck is not technical. It is political and commercial. The question is whether the regulator has the will to mandate genuine access at reasonable prices, and whether the political environment supports that kind of intervention against powerful incumbent banks.
The deeper insight here, and I think this is what Daniel was really driving at, is that we've been asking the wrong question. For years, the conversation has been: how do we make banks better at user experience? How do we force them to modernize their apps, extend their call center hours, improve their customer service? And the answer keeps coming back: they won't, or they can't, or they'll do it just enough to check the regulatory box. The better question is: how do we separate the UX layer from the settlement layer so each can be optimized independently? Let the people who are good at infrastructure run the infrastructure. Let the people who are good at building products build the products. The current system forces one organization to be good at both, and the result is that neither is as good as it could be.
The Cal call center closing at four PM is the perfect symbol of this. That's not a technology problem. It's not a regulatory problem. It's an incentive problem. Cal has no incentive to keep its call center open later because the cost of doing so is real and immediate, and the benefit, in terms of retained customers or new business, is diffuse and uncertain. In a competitive market, the company that stays open until nine PM wins. In an oligopoly, nobody wins because nobody has to compete.
Let's leave with one more question, and then we'll wrap up. If banks become invisible utilities, who ensures they maintain their infrastructure properly? The risk is a race to the bottom on settlement reliability. If the bank is no longer customer-facing, the pressure to keep the plumbing in perfect working order becomes indirect. The fintechs would scream if settlement slowed down, but they'd be screaming at a utility they can't easily switch away from. This is not a theoretical concern. We've seen it in telecom, where MVNOs sometimes struggle with network quality issues that are entirely outside their control.
Which points toward the logical endpoint of this model. A banking operating system. A standardized middleware layer that any fintech can plug into, maintained by a consortium or a regulator, with guaranteed service levels and transparent pricing. That's not where we are. That's not even close to where we are. But if you follow the decoupling logic all the way down, that's where it leads. The banks become pure infrastructure providers, and the interface layer becomes a competitive market with low barriers to entry.
The four PM call center closure is not just bad customer service. It's a signal that the people who control the plumbing have no incentive to care about the taps. And until that incentive structure changes, either through competition or through regulation, the taps will remain exactly as frustrating as they are today.
Now: Hilbert's daily fun fact.
Hilbert: In the seventeen eighties, French naturalist Pierre Sonnerat documented a whale call off the coast of French Guiana that local fishermen called "la flûte des caraïbes," the Caribbean flute. The name was a direct translation of the Kalina people's term for the sound, which they believed was not a whale vocalization at all, but the voice of a drowned musician whose instrument kept playing beneath the waves.
I don't know what to do with that.
A drowned musician playing a flute inside a whale. That's going to sit with me.
This has been My Weird Prompts. Thanks to our producer Hilbert Flumingtop. You can find every episode and the full archive at my weird prompts dot com. If you want to send us a prompt like Daniel did, email the show at show at my weird prompts dot com. We'll be back with another one soon.
