top of page

ISO 20022: Why Your Core Must Speak the New Language of Payments

  • Writer: WAU Marketing
    WAU Marketing
  • Feb 17
  • 4 min read

Updated: Jun 23

The world's language of payments changed in November 2025. And many cores still speak the old one—translating badly, losing words in every sentence.


It's not a comfortable metaphor; it's what happens technically. On November 22, 2025, the coexistence period SWIFT had opened in 2023 came to an end: from that date, cross-border payment messages between institutions are exchanged exclusively in ISO 20022, as SWIFT confirmed. The old standard—the MT messages that held up banking for decades—stopped being an option for those instructions. The transition was fast: by the end of 2024 barely about a third of payment-instruction traffic traveled in the new format, and by the close of coexistence it was around 80% of daily traffic, according to SWIFT.


What ISO 20022 is, and why it isn't just "another format"


ISO 20022 is a financial messaging standard built on structured, rich data. Put simply: where the old format crammed a payment's information into narrow text fields with hard character limits, ISO 20022 organizes that same information into an orderly, much larger structure. The full name of the parties, the legal identification of each entity, the purpose of the payment, the remittance details: it all fits, all tagged, all machine-readable.


That isn't cosmetic. That data changes three concrete things:


  • Reconciliation. With richer, more consistent information captured at the source, automated payment matching raises its hit rate. Fewer items a human has to chase by hand.

  • Compliance and fraud prevention. The structured fields—legal entity identifiers, payment purpose codes, the ultimate debtor and creditor—give your compliance team real context for every transaction. The U.S. Federal Reserve says it plainly: the ISO 20022 format enables more intelligent risk and fraud mitigation and provides richer data that facilitates sanctions and anti-money-laundering compliance, per its official Fedwire migration announcement.

  • End-to-end automation. Less manual intervention means payments that process themselves, faster and with less error.


This isn't only a European thing


The big payment infrastructures have already migrated: the euro system in March 2023, with the launch of the Eurosystem's new T2; the UK's in June 2023, when the Bank of England's CHAPS and RTGS moved to the standard; and the United States' high-value transfer network—which settles more than $4.7 trillion a day—in July 2025, according to the Federal Reserve. And it's not an isolated effort: the Bank for International Settlements' payments committee defined a set of harmonized data requirements for cross-border payments and asked participants to align by the end of 2027 at the latest, as the BIS details. Its own monitoring found that close to 79% of the world's real-time gross settlement and fast-payment systems have already implemented ISO 20022 or have concrete plans to do so, according to the BIS.


The Mexican case: two worlds your core has to connect


Here it's worth getting one thing straight that the press tends to blur. In Mexico you have to separate two planes:


On the cross-border plane, Mexico is already in. The Bank of Mexico adopted ISO 20022 for its message exchange with CLS, the international currency-settlement infrastructure, live since December 2, 2024, according to Banxico's own statement. Any Mexican institution with payments abroad already has to handle the new language.


On the domestic plane, SPEI runs on a proprietary Banxico protocol, distinct from ISO 20022, and has no full migration to the standard scheduled, as the Bank of Mexico documents. This creates the real challenge for your core: living between two worlds. Receiving an international payment in ISO 20022, loaded with structured data, and having to converse it with your domestic rails without losing that information along the way. That's exactly where a poorly prepared core fails.


The legacy core problem: truncation


The old format used fields with strict limits. When a rich ISO 20022 message has to pass through a legacy engine that wasn't designed to store those extended fields, the inevitable happens: the information gets truncated. Cut off. The full name gets clipped, the structured address gets crushed into a free-text field, the compliance data is lost. You received a payment with all the information and your system threw it away because it had nowhere to put it.


And the clock is ticking. The next hard milestone is November 2026: unstructured addresses will no longer be accepted in cross-border messages; they'll have to be sent in structured format, with street, city, and country in separate fields, according to SWIFT. The specialist consultancy RedCompass Labs found that close to 44% of banks are not on track to meet that deadline, and that three in five report gaps in their core banking systems precisely for handling structured addresses, per its research reported by Treasury Today. Anyone who papers over the problem with a translation layer that still truncates is just kicking the can.


How we see it at WAU


At WAU we prepare cores that speak ISO 20022 natively: that store the full structured data instead of crushing it, that translate between the international world and your domestic rails without losing information, and that reach the 2026 deadlines without fragile patches. Your core shouldn't be the link that loses words in every payment.


If you want to know where your current architecture truncates data and what it takes to speak the new language without losses, let's talk. We'll help you run the diagnosis. 👉 Book a conversation with our team.


Sources


Comments

bottom of page