EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
For the last few decades, treasury functions have relied heavily on spreadsheets, but growing complexity, fragmented data and the need for faster decisions are exposing this model’s limitations. In this episode of the EY India Insights podcast, Hemal Shah, Partner, Consulting Services, EY India summarizes Agentic AI playbook for CFOs and treasury professionals, which can help them build more intelligent, connected and trusted systems. He discusses Agentic AI’s adoption across cash flow planning, foreign exchange risk management and decision support, while emphasizing the importance of strong data foundations, governance and human oversight.
Key takeaways
Agentic AI transforms treasury by connecting fragmented data, identifying priorities and recommending actions, while preserving human judgment for important decisions.
Treasury functions must move beyond spreadsheets, as data silos, manual reconciliation and outdated reporting may restrict speed and strategic decision-making.
AI pilots may fail when organizations automate before establishing consistent, governed data that treasury professionals can trust.
By automating routine treasury tasks, Agentic AI enables professionals to focus on liquidity strategy, risk management and higher-value decisions.
Responsible adoption requires a phased roadmap. Assess maturity, prioritize low-risk use cases, establish governance, validate recommendations and then scale responsibly.
Successful treasury AI adoption requires strong governance, measurable business outcomes and a phased approach that builds trust over time.
Hemal Shah
Partner, Consulting Services, EY India
For your convenience, a full text transcript of this podcast is available on the link below:
Welcome to the EY India Insights podcast. I am Pallavi, your host for today. In this episode, we explore how Agentic AI is helping treasury functions move beyond spreadsheets toward more intelligent, connected and trusted operations.
Hi Hemal, thank you for joining us today and welcome to the podcast.
Hemal
Hello, everyone. Thanks for the opportunity.
Pallavi
Treasury teams have relied on spreadsheets for decades. Why is the model becoming increasingly difficult to sustain?
Hemal
Spreadsheets have served treasury teams quite well and they are still serving but the complexity, operational risks and speed of today's business environment have exposed their limitations.
A mature treasury function may operate between 50 to 100, either disparate or interconnected spreadsheets, covering various functions such as banking, bank account management, cash positioning, foreign exchange exposures, balance sheet exposures, cash flow exposures and hedge management. They may also cover investment management, debt management, short-term, medium-term, long-term debt, bank-led debt, capital market related debt, covenant monitoring, regulatory reporting and internal Management Information System (MIS).
Each of these spreadsheets can become a separate data silo, creating challenges around version control, manual reconciliation and risk of human error. Treasury professionals also spend a significant amount of time collecting data from different systems and spreadsheets, validating figures and preparing or customizing reports. By the time these reports reaches the decision makers, the information may already be outdated.
Therefore, moving beyond spreadsheet is not simply about adopting a more modern tool. It is about creating an integrated treasury environment that provides data/information, more insights, improved visibility and allows professionals to focus on strategic decisions rather than on manual coordination.
Pallavi
How is Agentic AI different from the conventional treasury automation and where can it create value?
Hemal
Traditional treasury management systems generally follow predefined rules to complete a transaction execution lifecycle and its repetitive tasks. A transaction life cycle will typically record and report the transaction or maybe account for it as well. There is a lot which is happening before and after the transaction. Agentic AI goes further by applying reasoning across this workflow, responding to new information and recommending actions within clearly defined governance boundaries.
For example, in forex risk management, Agentic AI can combine real time exposures with market data, assess multiple hedging scenarios based on past historic performance, do risk quantification - a simplistic or complex Value at Risk (VaR), quantify these exposures and help generate various types of scenarios, documentation and audit trail. These activities were disparate in nature, disaggregated in nature, but they can be combined using help of agents.
Let us take another example, of cash flow planning. Agents can consolidate balances across multiple bank accounts in different geographies and currencies, identify emerging patterns, historical insights and flag potential funding gaps for seven-day period, 30-day period, 60-day period in advance. So, the real value lies in connecting data, integrating this analysis, insights and helping with decision support.
Rather than merely presenting information in its static form, an Agentic AI can help treasury teams identify what requires their immediate attention and recommend the next set of appropriate actions. However, human judgment remains critical, particularly for material financial decisions and exceptions.
Pallavi
Pivoting towards AI pilots, why do many AI pilots in treasury fail to progress beyond the experimentation stage?
Hemal
As per our experience, one of the most common reasons is incorrect sequencing. Organizations sometimes implement AI tools before addressing the quality, consistency and governance of the underlying treasury data.
When data remains fragmented across Enterprise Resource Planning (ERP) systems, treasury management platforms, bank feeds, contracts, emails and spreadsheets, AI may generate inconsistent recommendations. Now, this reduces user confidence and treasury teams often return to familiar manual processes. A governed treasury data lake should therefore be established before Agentic AI is scaled. It brings structured and unstructured information into a single trusted environment, transforms it appropriately and provides agents with timely and reliable data.
It is like doing Quality Assurance (QA) and Quality Control (QC )function around your treasury processes. The important point is that an AI solution cannot compensate for weak data foundation. Organizations need to begin with data readiness – its ownership and governance. And once the foundation is established, AI pilots will have much stronger chance of producing dependable outcomes and earning the trust from the treasury professional or any other stakeholder.
Pallavi
Adding to the aspect of Agentic AI, does an agent reduce the role of treasury professionals, or does it enhance or change the way they contribute?
Hemal
Over here, I belong to a different school of thought. Agentic AI is more likely to enhance the capabilities of treasury professionals rather than replace their roles.
There will be huge reskilling across the front office, back office and mid-office roles. And in summary, new roles will also emerge. Today, considerable treasury capacity is consumed by gathering data, checking figures, reconciling information, producing routine reports, customized reports. When technology can handle more of this work, professionals can devote greater attention to better tasks like reevaluating a liquidity strategy, reforming their funding decisions, exploring various risk management options and engagement with business leadership.
Their role will increasingly involve interpreting AI generated insights, challenging the recommendations, addressing exceptions and applying commercial judgment. As I said, new capabilities will also be required, including data literacy, understanding how AI models operate and the ability to work within governance and control of these frameworks. Trust and accountability will remain essential. Treasury professionals must understand how a recommendation was produced and retain oversight on specific and significant decisions. The objective is not autonomy without control; it is more effective partnership with this specialized skill, expertise and technology.
Pallavi
According to you, what practical roadmap should CFOs and treasurers follow to adopt Agentic AI responsibly within the organization and their scope of work?
Hemal
Ideally, organizations should adopt a phased approach rather than a big bang approach or trying to automate the entire treasury function at once. They can begin with maturity assessment covering treasury data, the relevant technology processes, governance and workforce capabilities.
The next step is to select one high-impact, relatively low-risk use case. For example, tax forecasting or cash flow planning or cash reconciliation, where outcomes can be measured clearly. During the initial phase, an agent should ideally operate in advisory mode. This allows teams to compare recommendations with existing processes, validate output and identify exceptions before granting any execution authority.
I personally feel governance councils, approval workflows and audit logging should be established early. Wherever practical, agents can be connected to existing systems through Application Programming Interface (APIs), instead of replacing core infrastructure. As confidence grows, the organization can expand to additional workflow and establish a Treasury Center of Excellence to manage data pipelines, agent libraries and governance. Success should be measured through business outcomes such as faster turnaround times, improved forecasting, optimized liquidity, funding decisions, model accuracy and the overall blended time and cost optimization. That is how a business case can be evaluated.
Pallavi
Thank you, Hemal. Now that brings us to the end of this episode. Thank you so much once again for sharing all those valuable insights on how organizations can build more intelligent and trusted treasury operations.
Hemal
Thank you so much. It was a pleasure speaking with you all.
Pallavi
Thank you to all our listeners for joining us today. To learn more, please read the full article on ey.com. Until next time, this is Pallavi, signing off.
EY Treasury Consulting services help treasury teams & CFOs take steps to strengthen performance of their treasury function by transparency & internal controls.
Risk Consulting Services at EY helps identifying and managing risks across domains like Digital Risk, Enterprise Risk, Financial Services Risk, Actuarial and Risk Solutions to drive business success.
