Agentic AI Could Automate 80% of Routine Treasury Operations, EY Report Finds

Agentic artificial intelligence is poised to transform corporate treasury departments by automating up to 80% of routine coordination and processing tasks. A new report from EY indicates that this shift will allow finance professionals to pivot away from low-value administrative work, such as manual data validation and batch reconciliation, to focus on high-impact strategic activities including risk management and funding decisions. The findings suggest that while the technology offers significant efficiency gains, its successful implementation depends on companies first addressing fragmented data architectures.

Currently, treasury teams at mid-levels spend approximately 60% to 70% of their bandwidth on manual tasks, including gathering data from disparate systems and preparing reports. With more than half of global companies still relying on manual batch reconciliation, the potential for automation is substantial. By integrating information from enterprise resource planning systems, banking platforms, and other sources into a unified data lake, firms can leverage AI to achieve forecast accuracy of around 90% across 30-, 60-, and 90-day liquidity horizons.

The transition to an agentic AI-driven treasury model does not necessarily imply a reduction in headcount, according to the EY report. Instead, the objective is to reallocate human talent toward activities that generate greater business value. In a mature operating model, AI agents would handle the bulk of routine processing, while the remaining 20% of tasks—specifically those requiring nuanced judgment, complex risk assessment, and strategic oversight—would remain under the purview of human professionals.

This evolution is expected to help companies optimize their capital allocation by reducing the need for excessive liquidity buffers. By achieving higher forecasting accuracy, organizations can manage their cash positions with greater precision, effectively freeing up capital that was previously tied up in conservative, manual-driven safety margins. The report emphasizes that the path to this future is built through governed, automated workflows rather than wholesale replacement of human decision-making.

EY cautioned that the deployment of agentic AI is not a plug-and-play solution. Companies must first establish reliable data systems to ensure the AI operates on accurate, integrated information. The report highlights that a treasury data lake is essential for reliable decision-making, serving as a single source of truth that bridges the gap between various internal and external data sources. Without this foundation, the effectiveness of AI agents in managing complex treasury functions remains limited.

Governance frameworks are equally critical to the successful adoption of these technologies. The report advocates for an “advisory before autonomous” approach, where AI-generated recommendations are subjected to human review and approval before any action is taken. This is particularly important during the pilot phase, where companies are advised to avoid autonomous payment execution to mitigate operational risks while the AI systems are being calibrated and validated.

For treasury leaders and CIOs, the immediate priority is to identify and execute smaller, lower-risk use cases to build confidence in the technology. The report identifies cash forecasting, cash reconciliation, and the handling of exceptions in Know Your Customer (KYC) and Anti-Money Laundering (AML) processes as ideal starting points. These areas are characterized by high volumes of manual effort, making them prime candidates for automation that can deliver measurable outcomes in a relatively short timeframe.

To support this long-term transformation, EY recommends the establishment of a Treasury Center of Excellence (CoE). This body would be tasked with managing the data lake pipelines, maintaining workflow libraries, and enforcing data governance standards. As treasury functions become increasingly data-driven, the CoE will play a pivotal role in ensuring that the integration of agentic AI remains aligned with the broader strategic goals of the organization, ultimately building more resilient and agile treasury operations.

As organizations move forward, the integration of agentic AI is expected to redefine the role of the treasury professional. By offloading repetitive tasks to AI agents, teams can dedicate more time to complex financial analysis and long-term planning. The success of this transition will likely depend on the ability of leadership to foster a culture that embraces digital transformation while maintaining rigorous oversight of automated processes.

While the technology promises significant improvements in efficiency and accuracy, the industry remains in the early stages of adoption. Future milestones will likely involve the scaling of these pilot workflows into broader, more integrated systems. The unresolved question for many firms remains how quickly they can modernize their legacy infrastructure to support the data-intensive requirements of agentic AI, a challenge that will likely dictate the pace of adoption across the sector in the coming years.