The treasury function is facing an inflection point. As ISO standards reshape global payment rails and artificial intelligence (AI) capabilities transition from experimental to operational, treasurers face a choice: modernise payment infrastructure to capture real-time visibility and predictive intelligence, or maintain legacy systems that constrain business and competitive positioning.

For corporate treasurers, friction in the payments process has long been accepted as an operational reality, with manual processes, data siloes, and disparate legacy systems rendering corporate payments inefficient and opaque. But in an economy where capital shifts within seconds, payment latency has evolved from an operational inconvenience into strategic latency – time lost to internal questions about payment status, chasing colleagues and banking partners for information, and reconciling data across disconnected systems.

For the AI-first finance professional, this reactive posture is becoming unconscionable, as these inefficiencies increasingly constrain treasury’s capacity to serve as a true strategic partner to the business. Corporations that deploy technology to help eliminate this friction can unlock working capital, accelerate decision-making, and elevate treasury’s role as a strategic driver of enterprise agility.

“There is a saying in computer science that there are only two hard things – naming things and maintaining data consistency across distributed systems (caching) – both of these appear to apply to payments” says Luke Hammock, Managing Director, Global Head of Channel Products at Goldman Sachs Transaction Banking (TxB^TM).

“In payments, terms like ‘approved’, ‘complete’ and ‘settled’ all suggest a level of finality, but can mean different things to different players in the value chain, and may even be stored differently across distributed systems.”

The result is information ambiguity that can delay critical business decisions. “Once a payment is approved, the information the principal receives from their payment system or provider can be flaky and inconsistent,” says Hammock. “So, it can be difficult for the principal to be informed.”

The compounding cost of delay

Without the right technology foundation, payment friction can keep treasury teams in a cycle of operating reactively instead of proactively, creating cascading difficulties for forecasting and cash positioning. “AI can create opportunities to add interactivity and sophisticated simulations to your financial planning – but that doesn’t matter if you’re spending your entire day chasing down missing data,” says Hammock.

The opportunity cost extends beyond productivity. When treasury operates on batch-processed data and manual investigation, the entire enterprise loses agility. In an environment where competitors are making decisions in real-time, ten-minute batch delays compound into days or weeks of strategic disadvantage.

The convergence opportunity: three layers of transformation

The path forward lies not in incremental improvement, but in the convergence of three transformative forces that together enable treasury to operate at the speed of the AI economy.

The data foundation: ISO 20022

Payment friction may be commonplace – and in some cases, friction is a feature, not a bug. But treasurers have more opportunities than ever before to leverage technology to help reduce this unwanted friction.

ISO 20022 represents the first layer – rich, structured data that travels with the payment itself. As payment information moves along the value chain, companies can use structured remittance data to capture the purpose of a payment or the Stock Keeping Unit (SKU) being purchased within the payment message itself. Banks are required to pass this structured remittance data along to the beneficiary – giving companies the data they need to automate and streamline the reconciliation process, while eliminating time-consuming investigations.

“If information such as the payment’s purpose, beneficiary and invoice number is intrinsically part of that payment, you never have to go looking for answers to those types of questions,” Hammock explains. “That can be really powerful for a treasurer trying to move at the speed of the AI economy.”

Harnessing ISO 20022 is not a unilateral solve, and in some cases an entire ecosystem must update for the benefits to be fully realised. For example, there may be inconsistencies when using payment rails that are not yet using the data standard. Where treasury management systems (TMS) and enterprise resource planning (ERP) systems are concerned, technical upgrades may be needed to consume the relevant data and transmit it to banks. But the direction is clear: enriched data standards are the prerequisite for intelligent automation.

The delivery mechanism: cloud-based, API-first platforms

Against this backdrop, on-premise systems are increasingly giving way to agile, cloud-based platforms that harness the power of APIs – the second layer of transformation. The most progressive companies are leveraging APIs to merge hundreds, if not thousands, of discrete changes per day to production, with ‘real-time’ as the default answer for any new development.

“When I first came into this role, I really wanted to understand the power of real-time APIs in treasury,” says Hammock. “Often the answer was that today’s processes run on 10 to 15-minute batches, whereas APIs can perform an operation or give feedback immediately.”

Initially, Hammock felt that the difference between a couple of seconds and ten minutes wasn’t overly significant – but the strategic implications became clear upon deeper analysis.

“When you put those ten minutes together with numerous other ‘just ten-minute’ activities, you get a compounding of delays that can draw out decisions or agendas for days or weeks,” he says. “However, if you get an immediate response from the bank, the treasurer can complete that activity immediately.” The availability of AI is likely to increase our expectations for immediate information and action.

“APIs aren’t a panacea,” Hammock acknowledges. “But they are increasingly a necessary ingredient to the modern treasury stack.”

The intelligence layer: AI-ready operations

AI developments address the third layer: intelligent automation that transforms how treasury teams interact with data and systems.

“Incorporating AI into your workflows allows you to talk to machines in a way that was previously impossible,” says Hammock. “Teams can bypass clunky implementation exercises by using AI – whether that’s by building their own TMS, or handling workflows that would otherwise require IT involvement.”

Together, these three layers create a compounding effect: ISO 20022 provides the rich data that feeds AI insights; APIs deliver those insights in real-time; and AI transforms how treasury teams act on information. The result is a treasury function that anticipates rather than reacts; that automates reconciliation rather than investigates it; and that operates as a strategic business partner, rather than a back-office function.

The strategic imperative: taking the next steps

For treasurers looking to eliminate friction and capture strategic value, Hammock emphasises the importance of understanding what an API is and the value it can bring. “For example, what’s the difference between a synchronous API request and an asynchronous request? Treasurers may require some basic education that their banking partner can provide,” he remarks.

Where AI is concerned, Hammock’s advice is clear: “Don’t wait for AI to become perfect. It’s not too soon to start seeing how you might use it to spot problems, and maybe even fix them. And don’t be shy – bring your bank partners in to help you solve the problems you’re looking at.

“Some of the most progressive treasurers aren’t trying to do everything with AI immediately – but equally, they’re not holding back.”

Action points for treasurers

The transformation to real-time, AI-ready treasury operations requires deliberate action across four strategic dimensions.

1. Audit your payment friction. Identify the main sources of friction in your workflows, and pinpoint how much time is spent on reconciliation and payment investigations.

2. Assess your ISO 20022 readiness. Engage with your banking partners to understand their timeline for ISO 20022 adoption, and how you can leverage enriched data fields.

3. Explore API-first integration. Investigate how a cloud-native, API-first platform can integrate with your existing ERP and TMS to help create a unified view of global liquidity.

4. Monitor new tools and technology. Stay on top of industry developments and explore how innovative new technology could help to reduce friction.

To learn more about Goldman Sachs Transaction Banking visit gs.com/txb

Transaction Banking services are offered by Goldman Sachs Bank USA (“GS Bank”) and its affiliates. GS Bank is a New York State chartered bank, a member of the Federal Reserve System and a Member FDIC. For additional information, please see Bank Regulatory Information.