Spreadsheets still eat up 70% of treasury teams’ time as agentic AI promises sharper forecasts

Treasury teams are still spending as much as 60%-70% of their time on manual and low-value tasks, leaving little room for strategic work, according to a new report by EY India.

The report, An Agentic AI Adoption Playbook for CFOs and Treasurers, said spreadsheet-driven treasury operations can see forecast variances of more than 20%, making it harder for companies to manage liquidity effectively.

Agentic AI could change that. EY said AI-enabled treasury models can improve cash forecast accuracy to as much as 90% across 30-, 60- and 90-day liquidity horizons, allowing companies to make faster decisions around cash and liquidity.


Despite growing investments in treasury technology, spreadsheets remain deeply embedded in day-to-day operations. A mature treasury function can have 50-100 interconnected spreadsheets covering cash positions, foreign exchange exposure, investments and regulatory reporting, the report said.

More than half of corporates globally also continue to rely on manual reconciliation, according to EY. This creates inefficiencies and increases operational risks, while also highlighting the scope for workflow automation and AI-led transformation.

“Many treasury teams continue to rely heavily on spreadsheet-based processes at a time when organizations are seeking greater visibility, agility and control,” Hemal Shah, partner, risk consulting, EY India, said.

Agentic AI can help move treasury operations from a reactive to a more predictive model, Shah said, but companies will need strong data foundations, governance and clearly defined workflows to make it work.

EY said workflow transformation should be the first step for companies looking to deploy agentic AI in treasury. Organisations that have already introduced digital breaks and workflow automation are seeing 80%-90% auto-match rates in reconciliation, according to EY India analysis.

The report identified cash forecasting, reconciliation and KYC/AML exception handling as some of the most promising early use cases for agentic AI.

Cash forecasting could deliver the biggest business impact, EY said, while AI agents could also handle 70%-80% of routine KYC/AML exception cases with full auditability. This could allow treasury and risk teams to focus on more complex tasks.

But getting there will require companies to first fix their data architecture. EY recommended building a treasury data lake that acts as a single source of truth by bringing together data from ERP systems, banking platforms, contracts, emails and market information.

The report also recommended setting up a Treasury Center of Excellence to oversee data pipelines, workflow libraries and governance frameworks.