Ran Grushkowsky is CEO of MassPay, a global payout orchestration platform specializing in instant, real-time payouts worldwide.
In an article published in April, I wrote that Net-30 was dead and that speed was the new competitive weapon in payments. Now that the industry and technology are starting to catch up on speed, the next challenge is already in view.
JPMorgan’s 2026 outlook for cross-border payments puts it plainly: “Speed is no longer a differentiator: it’s a baseline expectation.” Banks and processors are investing accordingly. KPMG research cited in that outlook found 93% of financial institutions are modernizing their payments infrastructure.
Plenty of platforms are still catching up, and they should, but speed is now the price of entry, not the prize. The prize is whatever was hiding behind speed. In payouts, that is accuracy.
The sender measures a payout by how fast it leaves. The payee measures it by whether it arrives. Those are not the same event, and solving this gap is what’s next after speed has already been settled.
Why ‘Fast’ Does Not Equal ‘Landed’
Most of the time that payments fail is due to incorrect data, not because the rails broke. Swift’s analysis of payment exceptions on its network found that 72% came down to formatting errors, account issues and invalid data. Not settlement outages or correspondent failures, but fields.
A decade of investment went into the rails, and almost none into the data that rides on them. The result is an industry that can move money across the world in seconds, but still cannot reliably confirm that the name on the instruction matches the name on the account.
Retail solved this years ago. Type a shipping address into any competent checkout, and it corrects the street name, flags the missing apartment number and refuses the impossible postal code. All before you pay, while you are still there to fix it. Nobody ships the package first and discovers the bad address three weeks later. Payouts still do exactly that.
The Cost Is Split Very Unevenly
Ask a finance team what a failed payment costs, and you will get a small number.
Research conducted by Capgemini Invent for LexisNexis Risk Solutions, surveying 400 payment executives, found an average per-payment fee of $12.10 for rejected or repaired payments. Beneficiary name and address details were the single most common source of delay or failure, cited by 21% of respondents. And 72% of those executives were still checking that data by hand.
Twelve dollars is a rounding error, and it is only the fee, before anyone counts the labor behind it. That is precisely why this stays broken: The party with the power to fix it barely feels it.
Now, ask the payee. A contractor in Nairobi who invoiced three weeks ago sees a payment marked “sent” that is not in her account. She does not know whether it bounced, whether it is sitting at an intermediary or whether it is gone. She cannot escalate to a correspondent bank she has never heard of. She waits, then files a ticket, and 11 days later the money returns to the sender minus fees. Then someone asks her to confirm her account details again, as though the error were hers.
She absorbs the entire cost of a $12 problem. Multiply that across a supply side, and it stops being an operations line item. Accuity, now part of LexisNexis Risk Solutions, estimated that failed payments cost the global economy $118.5 billion in fees, labor and lost business in 2020. Sixty percent of the organizations surveyed said they had lost customers because of them.
A Question Regulators Have Already Settled
While the industry debated rails, regulators quietly decided the accuracy question.
Under the EU’s Instant Payments Regulation, Verification of Payee became mandatory for euro-area payment service providers on October 9, 2025, with providers outside the euro area following by July 9, 2027. And the regulation does not merely require the check. It requires that the check be offered to the payer free of charge.
Meanwhile, in the U.K., where the Payment Systems Regulator has directed some 400 firms to implement Confirmation of Payee, the scheme has run more than 2.5 billion checks since its 2020 launch and now covers almost 99% of Faster Payments and CHAPS transactions.
Two of the world’s most demanding payment jurisdictions looked at name-matching and concluded it is not a premium feature. It is a minimum standard, and nobody should be charged for it.
But notice what those mandates cover: domestic and regional transfers, inside jurisdictions with mature account directories. The corridors where verification matters most are the ones the rules do not reach. A payout into a mobile money wallet in a market with no central directory at all is nobody’s regulatory problem but yours.
What Accuracy Actually Asks Of A Platform
Accuracy asks three things of a platform, none of which is exotic:
• Validate before, not after. Account and name checks belong at onboarding, when the payee is inside your product and can correct a typo immediately. Not at disbursement, when they are asleep and the money is already moving.
• Make verification free internally. The moment it carries a per-check fee, someone reasonable decides to sample rather than check everything and the failures concentrate in exactly the low-value, high-volume payouts where the payee is most exposed.
• Own the failure. When a payout does not land, that is a defect in your product, not a data-entry mistake by the person waiting on the money. At MassPay, we stopped charging for payouts that fail. It sounds like pricing. It is really a forcing function, because it puts the cost of inaccuracy on the only party in the chain who can engineer it away.
The Bottom Line
The last five years of this industry were about making money move faster. The next five are about making sure it lands.
A payment that fails in four seconds is not faster than one that takes three days and works. It’s a disappointment delivered more efficiently. Speed gets money to the edge of the payee’s life. Accuracy is what gets it across.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
