For more than a decade, treasury transformation has largely been focused on visibility. Organisations invested heavily in digitisation to improve cash visibility, automate reconciliations, strengthen forecasting capabilities and provide treasurers with more timely information.
That investment delivered meaningful value. Treasury teams today generally have a far clearer view of their liquidity positions than they did a decade ago. Yet visibility alone is no longer enough.
The next chapter in treasury's evolution will not be defined by how much information organisations can access, but by how effectively they can turn that information into action.
As artificial intelligence (AI) matures, and particularly as agentic AI gains traction, treasury has an opportunity to move beyond monitoring and reporting towards decision support, intelligent workflows and faster execution, while maintaining the governance and control that the function demands.
The question facing corporate treasurers is therefore changing. It is no longer simply: How do we gather better information? Increasingly, it is: How do we move from information to action faster, without compromising oversight and accountability?
A new era for treasury decision-making
Treasury teams operate in complex environments where decisions often need to be made quickly and with incomplete information.
Each day presents a constant stream of questions. Where is our cash? What has changed overnight? Which entities require funding? Are exposures within approved limits? Which transactions require escalation? Where is liquidity sitting idle?
Historically, answering these questions required treasury professionals to assemble information from multiple systems, analyse data manually and draw conclusions themselves.
Agentic AI introduces a different possibility.
Rather than simply presenting information through dashboards and reports, AI-powered systems can analyse large volumes of data, identify exceptions, recognise patterns and generate recommendations aligned to predefined treasury policies and risk frameworks.
In practical terms, this means treasury professionals may spend less time gathering information and more time exercising judgement on the issues that matter most.
The role of the treasurer shifts from information processor to strategic decision-maker.
When intelligence becomes actionable
The real opportunity lies not in replacing treasury professionals, but in augmenting their capabilities.
Imagine a treasury function that starts each day with a prioritised briefing. The system highlights excess liquidity, identifies entities at risk of overdraft, flags unexpected movements in cash balances and surfaces potential foreign exchange exposures approaching policy limits.
Rather than searching for issues, treasury teams are presented with them.
Instead of reviewing multiple dashboards to understand the situation, they receive context, analysis and potential courses of action.
This progression from data to insight, from insight to recommendation and from recommendation to execution represents one of the most significant opportunities AI presents to treasury.
Importantly, it also aligns with a broader shift taking place across financial services. Clients increasingly expect technology to provide not just information, but meaningful guidance that helps them make better decisions.
The future of transaction banking will be built on better decisions
For banks, the implications are significant.
Corporate banking platforms have traditionally focused on access to information and transactional capability. Clients could view balances, make payments, download reports and manage accounts.
Future expectations are likely to be far more demanding.
Treasurers will increasingly expect systems that can answer more sophisticated questions. What changed? Why did it change? What are the implications for liquidity?
This creates a new opportunity for transaction banks to evolve from providers of information to providers of intelligence.
The institutions that succeed will be those that combine deep treasury expertise, strong connectivity and advanced technology to help clients navigate increasingly complex financial environments.
Success will depend not only on technological capability, but also on trust.
Treasurers must have confidence that recommendations are transparent, explainable and aligned to their organisation's objectives, rather than the interests of any individual service provider.
Governance will determine success
As exciting as the technology is, the future of treasury will not be defined by autonomy alone.
Treasury is responsible for some of the most critical financial decisions within an organisation. Cash, liquidity, financing and risk management functions require robust governance and clear accountability.
Not every decision should be automated .
A useful way to think about agentic AI is as a spectrum. At one end, systems monitor activity and generate alerts. Next, they provide recommendations. Beyond that, they may prepare instructions for review and approval. Only in carefully controlled circumstances should they execute predefined actions within clearly defined limits.
The appropriate level of autonomy depends on the nature of the activity, the amount of money involved, the level of risk and the clarity of the governing policy.
Simple, repeatable and highly structured processes may be suitable for greater automation . Decisions involving material judgement, significant exposure or heightened fraud risk will continue to require human intervention.
The objective should never be maximum automation.
The objective should be responsible automation.
Human judgement remains indispensable
Despite significant advances in AI capabilities, treasury remains fundamentally a human discipline.
Technology can identify patterns, process data at scale and apply rules consistently. It cannot fully replicate commercial judgement, contextual understanding, accountability or strategic thinking.
Indeed, one of the greatest risks may be overreliance on systems that are usually correct.
Effective governance requires more than a human sign-off. Treasury professionals need access to the assumptions, methodologies and alternatives behind recommendations. They need the ability to challenge conclusions, override decisions where appropriate and understand how outcomes were generated.
The most successful treasury functions of the future will combine the strengths of people and technology rather than treating them as substitutes.
Treasurers will increasingly spend their time asking higher-value questions: Are the assumptions sound? What are the risks? How would outcomes change under different scenarios? Is the recommendation commercially sensible?
These are questions that remain firmly in the human domain.
The real opportunity ahead
There is a temptation for organisations to begin their AI journey with technology.
Treasury should start somewhere else.
The better question is where decision-making is slowed down by fragmented information, repetitive analysis or manual investigation. Which processes consume disproportionate amounts of time? Which exceptions are repeatedly reviewed? Which policies are clear enough to be codified?
In many cases, the answer may not be agentic AI at all. Sometimes a better process, stronger integration or straightforward automation will solve the problem more effectively.
The organisations that derive the greatest value from AI will not necessarily be those deploying the most sophisticated technologies. They will be those applying the right technologies to the right problems, supported by strong governance and clear accountability.
The future of treasury is therefore not about handing decisions over to machines.
It is about creating an operating model where treasury professionals are empowered with better information, better insights and better recommendations, enabling them to make decisions faster and with greater confidence.
Ultimately, the question is not whether AI will play a role in treasury. It already does.
The real question is how treasury teams can combine human expertise and AI to make better financial decisions, faster, while remaining firmly in control. That is what agentic AI truly changes.
Share
