Flexibility as a Requirement: What Instant Payments, ISO 20022 and Stablecoins Mean for Bank Payment Architecture

By Sławomir Paśko, Founder and CEO at Neti
In 2023, the payments industry agreed on a clear deadline. From November 2026, cross-border payment messages sent over Swift would have to carry the payer's and payee's addresses in structured form, with at least the town and country in separate fields rather than one line of free text. In March 2026, about 65% of payment messages still did not meet this requirement (1). On 27 August, less than three months before the deadline, Swift deferred all payments changes planned for November. A new timeline is expected by December (2).
At Sibos a month later, Swift's Nicolas Stuckens asked whether the industry had focused "too much on message readiness rather than data readiness". Danielle Sharpe of Standard Chartered described the work as "an ongoing operations programme" and added: "This is a data programme" (3).
The lesson goes beyond one deadline. Change in payments has become continuous. The ability to absorb it is a requirement in its own right. Treating ISO 20022 as "just a format change" delays the real preparation (4).
Why each change costs so much
A payments team in a European bank has a long list on its roadmap: SEPA Instant, new ISO 20022 releases, stablecoins entering settlement discussions, tokenised deposit pilots and the digital euro. In many banks, each becomes a separate initiative with its own business case, budget and approvals.
Interbank settlement now has several answers. In August, HSBC and Standard Chartered made the first live tokenised deposit transfer on Swift's new ledger (5). In September, the ECB launched Pontes for DLT settlement in central bank money (6). Each works within its own network, with few bridges between them.
Research published before the deferral found that 44% of banks were not on track for the structured address deadline. More than 60% were delivering ISO 20022 changes through several parallel programmes rather than one (7).
What we heard from banks
In recent conversations with banks in Central Europe, three issues came up repeatedly. Transfers to some corridors still settle the next day or only within operating hours. Retail clients sending money abroad are moving to cheaper fintech apps. And the main obstacle to new payment products is often not compliance, but a shortage of people inside who can build and run them.
None of them asked for a new core system. They asked for a defined product with a clear corridor, service level and fee, delivered as a module next to their existing payment hub.
In my experience, the large global blockchain programmes in payments mostly did not work out. What is working now starts small: two banks, one currency, one corridor. Banks that wait for a global consensus before they start will wait a long time. Trust in this technology is growing from the bottom up, one working use case at a time. A small corridor that runs reliably for six months does more for adoption than any industry-wide roadmap.
Building next to the core
Some banking groups have stopped waiting for their legacy core to move faster. Core banking systems stay in place for around 20 years, and implementing a new one takes 4.7 years on average (8). A growing alternative is to build new capabilities alongside the core. Qivalis, a joint venture of 37 European banks including Poland's Bank Pekao, received an e-money licence from the Dutch central bank on 14 October. It plans to issue a euro stablecoin in the coming months (9). Not every bank will join a consortium, but the same principle can work inside a single bank.
What flexibility looks like in practice
In our experience, flexibility comes from an orchestration layer next to the payment hub: a control layer that makes each payment follow the process the bank defined, whichever rail carries it. The client does not need to see a token. The new rail works as a settlement layer, while the relationship, the account and compliance stay with the bank. Three design choices matter most:
Rails and providers are configured per corridor, so changing a provider is a configuration task, not a project.
Validations as replaceable modules, from sanctions screening to checking that a beneficiary bank is reachable on an instant rail.
One ledger and audit trail for every rail, so audit and compliance see a new rail the same way they see existing ones.
Questions worth asking
How long would it take to launch a new corridor with a new provider?
Can we change a validation rule without changing the payment flow?
Do we reconcile every rail in one place, with one audit trail?
If the answers are measured in quarters, the constraint is usually architecture rather than technology. Treating flexibility as a requirement is a smaller step than replacing the core, and it can start with a single corridor.
Sources
Swift, “ISO 20022 milestone for November 2026: Unstructured addresses to be removed”, 25 March 2026.
Swift, “Swift accepts community request to extend structured address migration for ISO 20022 payment messages”, 27 August 2026.
Finextra, “Sibos 2026: ‘ISO 20022 is the Sagrada Familia of the financial industry’”.
Asian Development Bank, “Solving Adoption Challenges of New Technology: The Case of ISO 20022 for Cross-Border Payment Messages”, ABMI Brief No. 13, October 2025.
HSBC, “HSBC and Standard Chartered execute first live tokenised deposit transaction on Swift’s blockchain-based ledger”, 20 August 2026.
European Central Bank, press release on Pontes, 21 September 2026.
RedCompass Labs, “Banks fall behind on ISO 20022”, 5 March 2026.
Bain & Company, “Disruption from Within: De-risking Banking Transformations at Speed”, March 2024.
Ledger Insights, report on the Qivalis e-money licence, 14 October 2026.


