Nearly all corporates are now exploring AI or automation, yet fewer than one in 10 treasury teams have successfully deployed it into daily workflows. Sam Hunt, CTO at MillTech, explores the gap between AI ambition and execution, and how it can be used effectively to enhance FX risk management.
For years, many treasury and finance teams treated FX risk as a manageable background issue, important, but rarely urgent. That assumption no longer holds.
FX volatility is rising, driven by tariff shocks, geopolitical disruption and diverging monetary policy, making risk management an increasing priority for CFOs and treasurers.
Yet research from MillTech found that while 94% of firms believe their FX strategy in 2025 was reasonably prepared for market conditions, 58% still reported losses.*
The challenge is no longer simply whether firms hedge, but whether they have an effective FX risk framework. Treasurers increasingly require operating models, workflows and infrastructure that deliver timely visibility, integrated insight and strong governance across the full FX lifecycle.
This is where artificial intelligence (AI) is beginning to reshape treasury operations in a meaningful way. With 99% of corporates now exploring AI or automation, the technology is moving FX management beyond backward-looking reporting and towards a faster, more proactive decision-support model.
Enhancing decision-making
Many corporates continue to struggle with fragmented service provision across their banking partners and platforms, a reliance on manual processes, and limited real-time visibility across their FX operations. These challenges make it difficult for treasurers to have clear and accurate views of realised and unrealised profits and losses at both organisation and currency level, access dynamic real-time data, and demonstrate best execution, impacting decision-making and ultimately financial performance.
AI’s value lies in its ability to process fragmented, multimodal information at a scale and speed that manual processes simply cannot match.
42% of firms are prioritising the adoption of AI for process automation, 40% for risk identification and 39% for risk management, highlighting its growing role in integrating and interpreting information from multiple sources to build a clearer, more timely picture of currency risk.
AI tools can simulate hedging scenarios and strategies, track hedging across complex, multi-currency portfolios and build a complete view of currency risk across exposures. Through this modelling, treasurers can better understand the impact of currency movements on cashflows, internal rate of return (IRR), liquidity and broader financial performance, altogether improving responsiveness and decision-making.
Managing operational risk
While AI should not replace human judgement, automation can significantly reduce the manual burden of data aggregation, analysis, and reporting, freeing up time for higher-value decisions.
Manual workflows remain widespread across the industry, with many firms still relying on phone calls, email instructions and file uploads for instructing FX transactions. These processes increase operational risk and heighten fraud exposure.
AI can add significant value by automating traditionally labour-intensive workflows. In practice, this means identifying exposures that require immediate attention, surfacing unusual counterparty concentrations, detecting liquidity inefficiencies, improving short-term cash forecasting, and enabling treasury teams to interrogate complex datasets using natural language rather than manual spreadsheet reconciliation.
AI can also strengthen fraud detection and prevention by learning normal payment patterns across amounts, payees, timing and geography, then flagging anomalies in real time to identify compromised credentials or insider threats.
This shift allows treasurers to build a more resilient, controlled and scalable operating environment, improving governance and reducing operational friction.
AI governance considerations
However, before deploying AI into FX risk management, firms must establish a clear governance framework that defines how AI will be used, who is accountable for its outputs, and what level of oversight is required. The more influence AI has on downstream financial decisions, the greater the need for a clearly defined model evaluation framework.
Corporates should also think carefully about operational resilience and third-party dependency risk, particularly where external AI providers are involved, as it can create new operational vulnerabilities at the same time firms are trying to reduce them.
Another important consideration is data usability. Many organisations already possess the information needed to improve FX decision-making, but it often sits fragmented across treasury systems, spreadsheets, broker portals, custodians and internal documents. AI is only as effective as the environment in which it operates, whilst model context protocol gives a powerful new interface for AI agents to engage with disparate data, the protocol is still early and should be introduced carefully.
Embedding AI in treasury workflows
Implementing AI in FX risk management should be approached as a phased operating model transformation, rather than the deployment of a standalone technology.
This begins with identifying high value, low risk use cases. Building proof of concepts (POCs) in these areas to demonstrate the value, and build support for heavier work to come later. Moving beyond POC takes a greater focus on governance and data. Over time, firms can selectively automate processes, creating rules-based tasks for AI while continuing to keep a human-in-the-loop.
AI’s potential for CFOs and treasurers also extends well beyond FX risk management alone. It can optimise cash forecasting and management, for example, reducing reliance on spreadsheets and manual inputs, and enabling treasury teams to model cashflow more accurately across multiple entities, currencies and time horizons.
This is becoming increasingly important as persistent inflation and heightened FX market volatility are making holding idle cash and managing FX exposure more costly. According to MillTech’s Global FX Report 2026, the cost of hedging rose by a mean of 67% in 2025, highlighting the need for greater operational efficiency and more intelligent treasury infrastructure.*
Through AI-enabled workflows, finance and treasury teams stand to benefit from stronger visibility, improved control, faster decision-making and reduced manual workloads at a time when treasury functions are under growing pressure to do more with less.
The most effective AI platforms are those that embed technology directly into the workflow itself, linking risk management, execution, treasury operations, and governance into one controlled environment.
While nearly all corporates are considering incorporating AI, fewer than one in 10 treasury teams have successfully embedded it into their daily workflows. As firms look to close that gap, the organisations most likely to succeed will be those able to use AI to demonstrate measurable improvements in visibility, execution quality, control, and operating efficiency within a governed and scalable workflow.
Important disclosures
MillTech is the trading name of Millennium Global Treasury Services Limited (“MGTS”), MillTechFX Americas Inc. (“MTA”), Millennium Global Investments Limited (MGIL) and Millennium Global (Switzerland) GmbH (“MGS”). MGTS is authorised and regulated by the Financial Conduct Authority (FRN 911636) and the National Futures Association as a Commodity Trading Advisor and Introducing Broker (NFA ID: 0529364). The registered address is 88 Wood Street, London, United Kingdom, EC2V 7QR. MTA is registered with the National Futures Association as a Commodity Trading Advisor (NFA ID: 0545635). MGIL is authorised and regulated by the Financial Conduct Authority (FRN 171039) and the National Futures Association as a Commodity Trading Advisor and Commodity Pool Operator (NFA ID: 0263119). The registered address is 88 Wood Street, EC2V 7QR, United Kingdom. MGS is a company registered in Switzerland with company number CHE 382.000.002. Registered address: Poststrasse 30, 6300, Zug.
The information herein is for information purposes only, intended for Professional Clients only, and does not constitute an offer, recommendation, or legal, tax, accounting or investment advice. The views expressed by individuals are their own and not those of the firm. Please refer to MillTech’s Legal and Compliance page for further regulatory disclosures and risk information, and the Research Disclosures page for details on the data provided herein.
