Modern payments under pressure: the technologies that hold up
By Nina Papazyan, Chief Product Officer, Clear Junction
In payments, adoption is easy to announce. Operating safely at volume is harder. The difference is evidence, exception handling, and accountability that stands up months later.
Institutions are under pressure to move money faster and across more markets, while remaining accountable for every transaction they process. That pressure is most visible in cross-border payments, where liquidity is required outside banking hours, settlement spans multiple rails and screening decisions must be made as transactions flow rather than reviewed later.
When a treasury team needs liquidity on a Sunday, an ops team is clearing exceptions at scale, and a compliance function needs an audit trail that still makes sense six months later, technology choices stop being theoretical.
The demand is familiar. What has changed is the expectation that emerging technologies can close these gaps quickly, even when operating models still depend on clear evidence, escalation paths, and accountable decision-making.
Some technologies reduce operational friction in settlement, liquidity, and control. Others create new reconciliation and oversight burdens as volumes grow. The practical test is whether they integrate cleanly into account structures and produce a defensible record of what happened, when, and why.
Technologies hold up in payments when they:
- Preserve finality and evidence through each step of the flow
- Scale exception handling without rebuilding the process around manual repair
- Keep controls inside the operating model, so screening, reconciliation and reporting stay aligned
What’s working, and what struggles to translate
Real-time payment rails are well established domestically, improving funds availability and visibility. The limitations become clearer when these rails extend across borders. Differences in settlement timing, operating hours and scheme rules create coordination challenges that speed alone does not resolve.
Credit arrives later than expected, balances sit in transit while confirmation is pending, and exception queues build up that teams have to clear manually.
The differentiator is orchestration: time-zone aware cut-offs, consistent reference data, and event-level status that finance and compliance teams can replay later.
Addressing these issues depends on how rails interact, how settlement and screening events are recorded, and how payments are reconciled when flows do not complete cleanly. Institutions that invest here focus less on throughput and more on finality, evidencing and control.
Stablecoins are gaining traction in specific treasury and settlement workflows where out-of-hours liquidity and confirmation timing create real cost.
Tokenisation is showing value where it removes operational breaks in treasury and settlement workflows. Sequencing improves, status becomes clearer, and fewer handoffs require teams to reconcile activity after the fact.
By contrast, technologies that sit outside core account operations create exposure as activity grows. Where compliance checks occur at the wallet level or after funds have already been received, teams end up reconciling the same flow across multiple systems. Screening outcomes, balances and settlement records are no longer visible in one place, which slows audit review and increases exception handling.
Institutions continue to rely on technologies that reduce operational breaks and make it clear how a payment moved from initiation through to settlement.
Where emerging tech struggles in production
- Split ledgers and parallel balances that force dual reconciliation
- Controls applied after funds arrive, creating downstream remediation work
- Audit trails spread across systems with inconsistent references
- Exception handling that grows faster than the ops team
AI in compliance: where it helps, and where caution remains
AI is now embedded in parts of transaction monitoring and screening workflows, primarily ahead of formal decision points. It is used to organise alerts, enrich transaction data and surface patterns for review, while responsibility for clearing or blocking activity remains with defined control functions.
Transaction screening operates with limited tolerance for error. Decisions have to be reviewed, explained and defended long after settlement. Review teams look for a clear record of why a transaction was flagged, which inputs influenced that outcome and whether the same result can be reproduced later.
Errors in screening decisions carry immediate downstream effects, including reporting obligations, customer remediation and supervisory engagement. Once funds have moved, there is limited scope to reverse the impact.
Models are tested against investigation and audit scenarios, not against scores that cannot be tied back to a specific screening decision. Where a clear, reproducible trail cannot be produced, AI is kept out of payment decision paths and limited to supporting investigation. In practice that means versioning, clear ownership, documented thresholds, and a record of what the model saw at the time of the decision.
Infrastructure for real-time, cross-border operation
As payment services extend across time zones and operating hours, infrastructure constraints become more visible.
Common pressure points include liquidity availability outside banking windows, reconciliation across multiple settlement rails and exception handling when transactions fail mid-flow. These issues intensify once services operate continuously and manual intervention becomes impractical.
Infrastructure that supports real-time, cross-border operation treats screening, reconciliation and reporting as integral parts of the payment flow. Evidence is generated as transactions progress rather than assembled later. Operating models assume continuous processing, including weekends and periods of market stress.
We see the same dynamic in multi-rail cross-border operations. Services evolve fastest when SWIFT connectivity, domestic rails, FX and reporting sit inside a single operating model, reducing handoffs between providers. In our own multi-currency SWIFT journey, moving from pooled incoming-only structures to named client accounts across 11 currencies with both pay-in and pay-out reflects that operational pull towards clearer ownership, reconciliation and control.
Stablecoins and programmable money: where traction is real
Stablecoins are already used by institutions for settlement and liquidity movement, particularly where traditional rails introduce delay or uncertainty. The risk profile depends less on the instrument itself than on how it is integrated.
Where stablecoin activity sits outside core account infrastructure, with separate wallets, balances and controls, teams must reconcile activity across multiple systems. Audit review then depends on piecing together records from different sources.
Where adoption is strongest, stablecoins are incorporated into existing account frameworks as programmable funding rails. Each incoming transfer is screened before credit, conversion into fiat follows defined controls, and crediting aligns with established reconciliation and reporting processes.
Consistent tooling remains essential. Address-level controls, transaction-by-transaction screening prior to credit and clear escalation points determine whether these flows remain governable at scale. The patterns that scale are consistent: transaction-level screening prior to credit, explicit hold points for review, defined conversion logic, and reporting that ties blockchain events back to account statements without manual reconstruction.
What defines usable innovation
Across payments, the technologies that endure are the ones that can run continuously, reconcile cleanly at volume, and stand up to audit and customer challenge months later. They fit inside existing account, treasury and control frameworks, and they reduce operational breakpoints as complexity increases.
Five production questions leaders ask
Finality. Liquidity location. Control gates. Exception ownership. Replayable evidence.
These five areas determine whether an innovation becomes a dependable capability or a source of operational drag.
From Clear Junction’s vantage point supporting institutions across multiple rails and time zones, the technologies that last are the ones that preserve evidencing and accountability under scrutiny as volumes grow and operating hours extend.