For decades, banks treated payments infrastructure as plumbing: essential, but not something worth reinventing while it still worked. That calculation is changing fast. What was once a back-office maintenance question has become a front-line competitive issue, and financial institutions that keep deferring modernisation are starting to notice the cost of standing still.
The clearest signal came from SWIFT itself. After a multi-year coexistence period, the global messaging network completed its migration to the ISO 20022 CBPR+ standard for cross-border payments in November 2025, ending the coexistence period for legacy MT payment-instruction messages in favour of ISO 20022-based messaging in favour of a data-rich structure built to support richer transaction detail, faster reconciliation and stronger fraud controls. Domestic real-time payment schemes are converging on the same standard, effectively setting a common language for how money moves globally. Banks that have not yet fully adapted their systems are facing increasing pressure to complete the transition.
The Technical Debt Banks Can No Longer Afford
Legacy payment architecture rarely fails outright. It degrades quietly, through mounting integration costs, longer release cycles and compliance workarounds that pile up faster than anyone budgeted for. According to the 2025 McKinsey Global Payments Report, the payments industry generated 2.5 trillion US dollars in revenue in 2025 from roughly 3.6 trillion transactions worldwide, a scale that leaves little room for institutions still running fragmented, siloed systems to compete on cost or speed. Fragmented architecture increases operational overhead at every layer, from separate systems for domestic and cross-border rails to manual reconciliation between legacy cores and newer digital channels, and it leaves compliance teams stretched thin trying to interpret regulatory requirements that were never designed with today's messaging standards in mind.
"Modernisation used to be framed as a multi-year strategic bet," said Ivaylo Bozoukov, a fintech executive who has spent much of his career building payments and digital wallet infrastructure across emerging markets. "That framing no longer holds. Banks are discovering that the cost of doing nothing, whether in compliance risk, customer attrition or the sheer difficulty of shipping new products, is now higher than the cost of the migration itself."
Why AI Ambitions Are Forcing the Issue
The pressure is not only regulatory. Banks are continuing to increase investment in generative AI, with payments operations among the areas being explored for automation and straight-through processingof banks are increasing generative AI spend in 2026, with payment operations identified as a priority target for automation and straight-through processing improvements. AI models, however, are only as useful as the data feeding them, and this is where legacy infrastructure becomes a genuine constraint rather than a mere inconvenience. Systems built around batch processing, inconsistent data formats and disconnected ledgers cannot easily support the real-time, structured data that modern AI tools require.
Ivo Bozukov argues this is the point many institutions are missing. "Banks are investing heavily in AI for fraud detection, cash forecasting and customer service, but too few are asking whether their underlying payments infrastructure can actually deliver the clean, structured data those models depend on," he said. Bozoukov argues, "You cannot bolt intelligent automation onto a system that still relies on batch files and end-of-day settlement. The infrastructure has to be rebuilt first, or the AI investment simply will not pay off."
Buy, Build or Partner
Not every institution is approaching modernisation the same way, and that divergence is becoming one of the more interesting stories in the sector. Smaller banks and credit unions, without the balance sheets to fund multi-year core replacements, are increasingly favouring platform-based partnerships that deliver modern capabilities, real-time rails, API connectivity, and ISO 20022 messaging without a full infrastructure rebuild. Larger institutions, by contrast, tend to pursue modular modernisation, replacing components incrementally around a stable core rather than attempting one disruptive migration.
Both routes carry real trade-offs. Platform partnerships move faster but introduce dependency on a third party's roadmap and risk profile. Modular in-house modernisation preserves control but takes longer and requires sustained investment across multiple budget cycles, a commitment that can falter when leadership priorities shift. There is no universally correct answer, only a decision that has to match an institution's size, regulatory exposure and appetite for change.
An Opening for Markets Without Legacy Baggage
For financial institutions across the GCC, East Africa and Latin America, this shift carries a different implication. In some emerging markets, institutions may face less legacy-system complexity than long-established banks in developed markets, potentially making newer payment architectures easier to adopt, decades of accumulated technical debt built on infrastructure never designed for today's requirements.
"Markets that are building payments infrastructure now, rather than retrofitting it, have a genuine structural advantage," Bozoukov said. "They can adopt ISO 20022-native systems and real-time rails from the outset instead of migrating away from something older. That is not a small advantage. It is the difference between building for the next decade and spending the next decade catching up."
That advantage will not last indefinitely. As global payment standards converge and customer expectations rise everywhere at once, the window for leapfrogging legacy infrastructure will narrow. For now, though, institutions in emerging markets have an opportunity that many of their developed-market counterparts would gladly trade for a clean starting point.
The Real Cost of Delay
Institutions still treating infrastructure modernisation as optional are effectively betting that the pace of change will slow down long enough for them to catch up later. Nothing in the current trajectory, from SWIFT's completed ISO 20022 migration to the growing dependence of AI strategy on structured payment data, supports that bet. How to sequence investment across real-time payments, messaging standards, API connectivity and cloud infrastructure remains a genuine strategic question for every institution. For many institutions, the strategic debate is increasingly shifting from whether to modernise toward how and when to sequence that investment.
Disclosure: The information in this article draws on industry research and executive commentary; readers should note that Ivo Bozukov and Ivaylo Bozoukov’s company operates in the payments and digital wallet space.