The National Payments Vision will be decided at the till

Two years into the National Payments Vision, the UK has a dated plan for almost everything except the place where most payments happen. The mechanism for that place went live in June.

In November 2024 the Treasury wrote down an ambition no British government had put in a policy paper before: that account-to-account payments should become “a ubiquitous payment method”. The word is doing a lot of work – a payment method is only ubiquitous when it works wherever a card does. In the UK that means the till.

The Payments Forward Plan that followed in February put dates on the machinery: the regulator merger, the open banking legislation this quarter, the FCA’s framework and a successor to Open Banking Limited. It mentions the point of sale once, in a list.

Here is the shape of the market the Plan is trying to move. In March this year UK acquirers processed 2.1 billion card transactions. Of those, 348 million were online. The other 1.8 billion happened in person, and 73% of the month’s card transactions were contactless taps. Open banking is eight years and one billion payments in; 40 million payments a month, almost all of them online. Pay by Bank has grown at the online checkout – the smaller of the two markets by a factor of five.

For us at Volt, bringing account-to-account payments to the terminal in the UK with Worldpay, now Global Payments, is the last piece of what we call feature parity with cards. Having spent several years of my career working on point-of-sale projects, I understand why the toughest piece of the vision is being tackled at the end. A terminal asks more of a payment than a checkout page does.

What a terminal asks of a payment

A payment at a till has to complete in the time it takes to tap. The merchant has to know there and then that it succeeded; the Bank of England’s infrastructure consultation calls this “certainty of fate” and puts it first among the conditions for account-to-account payments in store. The payment has to work in a supermarket queue and on a train with no signal. It has to be refundable at the same counter. And it has to do all of this for a transaction worth about £35, the UK average this spring. That leaves no room for a second screen.

Cards manage this because the hard part was done decades ago, and the customer carries the result in a wallet. With an in-person open banking flow, the customer selects Pay by Bank, scans a code, opens their banking app and approves the payment with their face or a passcode. Every in-store Pay by Bank deployment in the UK today uses a version of this flow. The most inventive is Costa’s: the customer pre-authorises the payment in an app before reaching the counter, then presents a barcode the till reads as a gift card. It works. It also tells you exactly what the terminal cannot do, because the payment had to be moved out of its way.

This is what UK Finance’s proof-of-concept programme found when it put account-to-account payments through live stores earlier this year: they work in controlled environments today, on workarounds that will not hold at scale, and without the payment guarantee a merchant gets from a card at the moment of sale. Its recommendation was more merchant pilots and a definition of point-of-sale-grade certainty. I agree with both.

The mechanism already exists, and it went live in June

A card works at a till because the customer’s authority was granted once, when the card was issued. Every tap since is checked against it, and the tap carries no decision of its own. Open banking had nothing like this until this year.

Then on 2 June the UK Payments Initiative’s commercial Variable Recurring Payments scheme went live, with banks covering about three-quarters of UK current accounts. The customer approves an agreement once in their banking app and sets a maximum per payment, a maximum per period and an end date. The agreement stays visible in that app, and the customer can pause or cancel it there. The scheme rulebook then allows the payment provider to initiate payments “with or without the Payer being present”, within those limits. Read that clause with a terminal in mind. It is the card’s authority, granted to a bank account.

The scheme was built for bills. Wave 1 covers utilities, financial services, government and rail. The second wave covers general e-commerce. It is due in the second half of this year and is waiting on a statutory instrument the Treasury has promised by December. Nobody has scheduled the till. UKPI’s managing director lists “over time, point of sale journeys” as a direction of travel, which is more than the Forward Plan says.

Meanwhile the policy answer to the till has moved to a different building. In November the Payments Vision Delivery Committee made account-to-account payments at the point of sale “the Committee’s priority” for the next-generation retail infrastructure. The Bank of England’s June consultation gave it a section of its own: certainty of fate, NFC or QR acceptance, a universal kernel for terminals and a scheme rulebook with minimum consumer protection. All of it correct, none of it dated. The consultation concedes that scaling account-to-account payments to high-volume stores “may require additional functionality”, and the infrastructure that would supply it is a multi-year build whose design closed for comment only last month.

So the UK has the till on the drawing board for the end of the decade and the mechanism for it in production since June. The question I want to put to the industry is whether we wait for the first or use the second.

Australia shows an agreement can carry a tap

We did not have to guess at the answer, because we had already built it somewhere else. Australia has had agreement-based payments on its real-time rail since 2022. PayTo is available on 95% of retail banking accounts. It lets a customer authorise an agreement in their banking app and manage it there. We are an NPP Identified Institution and since September we hold our own Australian financial services licence, so we build on that rail under our own permissions.

On it we have demonstrated an in-store payment end to end on a typical high street terminal. The customer approves a PayTo agreement once in their banking app, then holds their phone to the terminal. The tap is authorised against the agreement and the money settles to the merchant in seconds. There is no card in the transaction and no card scheme behind it. The customer did the authenticating once; the terminal did none.

At this point it is a demonstration: we have shown that the mechanism works, and that the speed and the experience at the till can match a card. How a country adopts it is a different question that nobody has yet fully tackled. Now that the Reserve Bank of Australia has cut interchange and removed card surcharging – taking roughly A$910 million a year off merchants’ card costs – I would argue it is the right moment to be clear about what the case rests on. It is not the per-transaction fee.

With Worldpay, now Global Payments, we are building the same capability for the UK on commercial VRP and working towards a pilot with merchants. It will be a pilot before it is anything else, and deliberately so. Its purpose is to test the question UK Finance put to the industry: whether an agreement-based payment can give a merchant the certainty of a card at the speed of a tap. If it can, the UK’s own infrastructure can carry the point of sale end to end with no card scheme in the loop.

Why this is a question of sovereignty

Governments and central banks want to own their payments infrastructure, because it is the foundation of economic activity. The UK owns its Faster Payments rail, its open banking standards and now an agreement scheme. It does not own the till. Visa and Mastercard carry more than 95% of UK card payments by value. On the regulator’s own finding their scheme and processing fees rose by at least 25% in real terms between 2017 and 2023. UK retailers paid £1.48 billion in card fees in 2024, more than double the 2019 figure. The Bank of England puts the average cost of accepting a card at about 0.6% of the transaction, with the smallest merchants paying over four times what the largest pay. The regulator has chosen transparency over a price cap. That is a defensible choice for a regulator. It leaves the structure of the market exactly where it was.

The countries that changed their market structure did not do it by regulating the schemes. India’s UPI carries 15 billion merchant payments a month. Brazil’s Pix did 80 billion transactions last year against 48 billion on cards. Both did it with a QR code and a scheme the central bank ran or mandated, at a consumer price of zero. In both the biggest loser was cash. Europe’s own attempts are more sobering: Swish is used for 2% of Swedish in-store purchases, Wero’s broad point-of-sale rollout has moved to 2027, and Poland’s contactless BLIK rides Mastercard’s tokenisation at the terminal. The one native account-to-account tap at scale in Europe is Bizum’s: live across three-quarters of Spanish terminals since the summer, with no volumes published yet.

The plain reading of that evidence is that a universal contactless card plus a phone wallet leaves the consumer nothing to gain, and the UK is exactly that market. I do not expect account-to-account payments to win the till on consumer demand. The case rests on what a merchant can measure: money that settles in seconds rather than at the end of a cycle, and a fee set by a scheme the UK controls. That is the sovereignty argument, and it is a merchant’s argument before it is a government’s.

Four decisions that do not need to wait for 2030

First, the scheme. A point-of-sale use case can be added to commercial VRP through the rulebook’s own change process, with its own merchant due diligence and dispute rules. That can happen long before the next-generation infrastructure exists. The Delivery Committee itself expects “product-level arrangements” for the point of sale to come from industry. Someone has to propose one.

Second, certainty of fate. Pay.UK proposed a Faster Payments rule clarification in January so that a merchant learns within the transaction whether a payment succeeded. It has no delivery date. With an agreement in place this is the only thing the terminal needs from the rail.

Third, the phone. On Android the NFC chip is open. On iPhone the picture is different: Apple’s UK platform for third-party NFC access has produced no UK products in 18 months by the CMA’s own count, and the CMA is consulting on remedies this autumn. In the EEA the same access has been free since July 2024. Until this is settled every iPhone in the country is outside the proposition, and no scheme rulebook can fix that.

Fourth, protection. Cards carry chargeback rights and Section 75 on credit purchases; account-to-account payments carry neither, and Which? found that only 13% of consumers know it. The gap is real, and it does not need a statute to close it. iDEAL has been live in the Netherlands for two decades and carries more than 70% of Dutch e-commerce, with no chargeback mechanism. A complaint goes to the retailer first and then to a binding dispute committee under the trust marks. The Dutch banks are only now adding purchase protection to its successor, with a 2028 target. Physical retail already works this way: the returns desk is the first instance for any dispute at a till, whatever the customer paid with. Where more is needed, the market is already competing to supply it. Visa and Mastercard both sell dispute layers for UK account-to-account payments, Amazon covers Pay by Bank with its own guarantee, and the Delivery Committee’s design of a minimum at the core and more at product level invites exactly that competition. A dispute layer from a card scheme keeps the scheme out of the payment and puts it in the rulebook, so the decision for a national scheme is which rulebook it adopts. That brings us back to the first item on this list.

What a cards market does once the card is optional

UK Finance expects cards to carry 65% of UK payments in 2035, almost exactly their share today. That forecast assumes the till stays where it is, and on the evidence of the last two years it is a sensible assumption: the recurring-payments scheme slipped a quarter and the point-of-sale infrastructure has no date at all. The burden of proof sits with account-to-account payments. That is why we are building towards a pilot and why I would rather the industry tested the agreement model at a hundred tills next year than designed the perfect kernel for 2030.

We have taken cards for granted for the better part of a century as the owners of the point of sale. For the first time the UK has written down that it wants an alternative there, and it has built the one mechanism that can supply it. The infrastructure programme will take years, as it should. The agreement scheme is live now. What does a cards market do once the card is optional at the till? The Vision asked the question in 2024. It can only be answered at a terminal with real merchants, and that is what we intend to do.