The proliferation of payment rails is creating an unusual problem for transaction banks. The industry is simultaneously improving conventional payment infrastructure while developing tokenised deposits, stablecoins and other forms of digital money, with artificial intelligence increasingly capable of determining how transactions move between them.

For corporate and institutional clients, however, the underlying requirement has changed much less. They still want payments to be faster, cheaper, predictable and transparent.

That gap between relatively stable client expectations and increasingly complex infrastructure could redefine the role of correspondent banks. Instead of providing reach principally through counterparties, currencies and countries, they may increasingly have to decide how transactions, liquidity and risk should move across multiple networks and forms of money.

Jonas Stepczynski, Head of Market Management, Institutional Cash Management at Deutsche Bank, sees orchestration as central to that evolution. His view is consistent with Deutsche Bank’s September 2026 work on the future of cross-border payments, which argued that banks increasingly need to coordinate activity across multiple networks, payment methods and forms of money.

Clients want the same outcome from a more complex system

Stepczynski identified three themes dominating payments discussions at Sibos: the movement of digital money from theory towards implementation, continued improvement of existing payment rails, and the shift of artificial intelligence from experimentation towards practical adoption.

The common factor, he said, is that each is creating investment decisions that banks have to make now.

Yet he argued that client demand itself has not fundamentally changed; the world around them has changed quite significantly.

Clients continue to seek faster, cheaper, more predictable and transparent payments. What is changing is the environment in which banks have to provide them.

With new payment platforms and forms of money emerging, clients increasingly have to consider not merely how to send a payment from one point to another but which rail they should use, which infrastructure investments will remain relevant several years ahead and how the different systems will ultimately interoperate.

This shifts part of the problem from execution to architecture. Stepczynski said clients will increasingly want banks to manage the choice of rail and interoperability across payment ecosystems while keeping the underlying complexity low.

Reach will extend beyond countries and currencies

Stepczynski said the industry will require both convergence and orchestration, that some convergence of standards will be necessary across the principal forms of digital money and their supporting infrastructures. Without it, he argued, the cost and complexity of adoption could outweigh the benefits.

But convergence is unlikely to produce a single system. Stepczynski expected multiple payment ecosystems, settlement assets and forms of money to remain, leaving banks to connect them. He sees the correspondent banking model changing.

He described this as a “redefinition of reach”, in which correspondent banking will extend beyond counterparties, currencies and countries to include access to payment ecosystems, settlement assets and forms of money.

Deutsche Bank has been exploring that changing landscape across conventional and tokenised infrastructure. Its May 2026 work on digital money describes an increasingly always-on financial system in which legacy batch-processing and cut-off structures coexist with stablecoins, tokenised deposits and other digital-money models.

Its more recent analysis of cross-border payments similarly envisages a multi-rail environment shaped by new payment systems, digital money, regulatory divergence and artificial intelligence (AI).

The strategic implication is that correspondent banking does not necessarily disappear as new rails proliferate. Its function may instead move up a layer.

Orchestration is more than intelligent routing

That role would involve considerably more than selecting the fastest or cheapest payment route. Stepczynski said orchestration requires banks to manage liquidity efficiently across different payment ecosystems and forms of money while maintaining a holistic risk-management framework across them. “It’s not only the access to the respective platforms,” he said.

A bank might technically connect to several networks yet still leave its clients with fragmented liquidity, separate controls and operational complexity. Effective orchestration would have to manage those elements together.

For clients, the desired outcome is almost the reverse of what is happening underneath. As infrastructure becomes more complicated, their interaction with it should become simpler.

Stepczynski expects clients ultimately to want a single entry point capable of managing transactions and liquidity positions while reaching multiple ecosystems without exposing them to the complexity behind those connections.

That gives large correspondent banks a potential advantage, but also raises the investment requirement. They have to maintain connectivity with existing rails while determining which emerging networks and settlement assets warrant integration, and then manage liquidity and controls consistently across them.

AI could become part of the orchestration layer

AI could help make that complexity manageable. Stepczynski sees applications in routing and client experience, with AI potentially smoothing many of the frictions clients encounter today.

The significance goes beyond automating an existing process. In a multi-rail environment, AI could eventually help banks determine how transactions should move across different networks while managing the complexity behind those choices.

Stepczynski did not describe such an autonomous orchestration model as a current Deutsche Bank deployment, and the distinction matters. His comments set out the direction in which the correspondent banking function could evolve, not a claim that all of those capabilities are already operating in production.

The immediate challenge is more fundamental: banks have to make infrastructure decisions while the eventual architecture remains unsettled.

For Deutsche Bank, that uncertainty does not diminish the role of correspondent banking. It changes what clients may expect it to provide. If payments fragment across more rails and more forms of money, reach will no longer be measured only by how many currencies, countries and counterparties a bank connects. It will increasingly depend on whether the bank can make those different systems behave as one coherent service for the client.