AI STRATEGY
At this stage, a practical AI strategy focuses on automating key processes to free up treasury’s time to improve performance elsewhere. The first step is to identify pain points that could be improved by the adoption of AI, then design and test a new process for potential efficiency gains before implementing the solution.
Identifying pain points
The 2025 AFP Treasury Benchmarking Survey identified a number of tasks that pose problems for corporate treasury departments. The most common responses were cash and liquidity forecasting, payments management and improving working capital, along with more general concerns regarding the challenges of automating manual processes and transforming treasury through the implementation of new technology.
At the leadership level, the survey identified some significant gaps between what senior treasury professionals consider to be important skills and their view of their own effectiveness at carrying them out. The identified gaps were all in areas that might be considered soft skills — communication, strategic thinking, collaborative skills and analytical skills — rather than technical treasury skills.
Using AI to manage pain points
It can be helpful to view AI as simply another tool to support the organization. Given access to the required data and designed to address specific issues within set parameters, AI can potentially identify patterns and trends and make reasoned decisions.
“AI is already helping treasury teams decide how long, how much and in which currency to invest,” said Britta Döttger, Head of Treasury at Roche.1
AI cannot replace some of the key value-adding roles of corporate treasury identified in the 2025 AFP Treasury Benchmarking Survey, such as strategic thinker, future planner, communicator and driver of change. Consequently, the general approach is to use AI to improve specific technical processes, especially those that consume disproportionate levels of management time. In turn, this will release time for treasurers to address the value-adding roles described above, and to address exceptions, including those activities that AI cannot deal with or that treasurers do not trust AI to manage. Over time, as treasurers become more comfortable with the technology, there will be opportunities to use AI, as Kelly suggested, to “reimagine multiple processes to create something meaningful.”?
“As Dana Laidhold at NASDAQ said, ‘We don’t deal with data — we deal with analysis.’ That shift only happens when you’ve automated the low-value tasks.” — Mike Richards, CEO The Treasury Recruitment Company
Developed targeted solutions
The focus on solving particular problems can add real value to corporate treasury departments. “The best examples are practical and driven by necessity: NASDAQ uses FX bots to scrape their balance sheet daily and execute trades — what used to take days now takes 20 minutes,” said Mike Richards, CEO, The Treasury Recruitment Company. “ASML built a cash forecasting tool in-house using just two people and open-source Python. It’s now 97% accurate and outperforms their manual forecasts. Twoday Group created AI tools with built-in rule sets, replacing the need for a TMS altogether in some cases. They built targeted solutions that addressed real inefficiencies.”
Herrera knew there was an error in the way the bank had calculated interest on some of Related’s many bank accounts. A simple AI query was able to identify the affected accounts, providing a quick resolution to what “would otherwise have been a time-consuming task,” she said.
Focus is the key. Success comes “not as much from using AI as it does from understanding treasury,” said Bob Stark, Global Head of Market Strategy, Kyriba. “You have to understand the process where you want to use AI. Then, you calculate how and/or why the AI process is better than that.”
Use efficiency gains to hone soft skills
In practical terms, efficiency gains often translate into working capital improvements. For example, more timely and accurate cash forecasting lets companies hold less precautionary cash, freeing funds to reinvest in the business or reduce borrowing costs.
Efficiency gains also lead to changes in the role of treasury. Over the last 20 years or so, the increasing use of technology in treasury has led to automation of core tasks, freeing time for treasurers to become more proactive within the business. AI represents the next level in this transition, offering greater opportunities for automation and deeper analysis of company data.
AI-driven automation frees treasurers to focus on highervalue skills, such as communication, which was highlighted as important in the 2025 AFP Treasury Benchmarking Survey. With more time and access to deeper analysis, treasurers can better support strategic planning and provide clearer evidence for change, making treasury more valuable to the organization as a whole.
PUTTING AI INTO TREASURY PRACTICE
While the potential benefits of AI implementation seem clear, determining how to proceed can seem daunting, especially for treasury departments with limited resources. For many, two key questions dominate. First, how can the department acquire the skills necessary to run one or more AI-powered processes? Second, how does the department ensure each AI-powered process provides a better outcome than the existing method?
“Technology doesn’t create value on its own, people do. Treasury teams need the skills and confidence to work alongside AI, applying judgment and business context to turn insights into better decisions.
— Ather Williams III Executive Vice President Head of Global Payments & Liquidity and Wholesale Digital Wells Fargo
There is value in experience. “The technology is a leveler: a young person might find the tech more instinctive, but more experience allows a deeper understanding of the context of any information and recommendations provided.”
— James Kelly Co-founder Your Treasury