Treasury is Betting Big on AI but Why Not All In Yet?
Artificial intelligence (AI) is quietly transforming treasury management into a more efficient and data-driven function. Tasks like cash flow forecasting, fraud detection, and credit risk assessment—once reliant on manual processes—are now being enhanced by AI, enabling treasury teams to operate with greater accuracy and speed.
KPMG’s AI in Finance report highlights this shift, with 64% of organizations piloting or actively using AI in treasury operations. The practical applications are clear: real-time data analysis allows for more precise financial forecasting, while pattern recognition in AI-driven systems helps identify irregularities and minimize fraud. Predictive analytics is further aiding credit risk evaluations, supporting better-informed decision-making.
Yet, adoption varies across functions. Treasury lags slightly behind financial planning (78%) and accounting (76%) but leads over tax operations, where adoption stands at 45%. This disparity reflects both treasury’s growing reliance on AI and the challenges of integrating advanced technology into legacy systems.
The financial returns from AI adoption in treasury are increasingly clear. According to the report, 92% of organizations using AI in finance report that their initiatives meet or exceed ROI expectations. This strong performance underscores AI’s value as a strategic tool rather than a mere operational enhancement.
Treasury teams leveraging AI report measurable improvements in accuracy and efficiency. Cash flow forecasting powered by AI reduces risks of liquidity mismanagement, while fraud detection systems minimize financial losses by identifying anomalies more quickly. Predictive analytics enables treasury teams to simulate market scenarios, providing valuable insights that were once time-consuming to produce manually.
Among companies classified as AI leaders, the returns are even more pronounced. These organizations dedicate 13% of their IT budgets to AI initiatives, a figure expected to rise to 17% within three years. This investment allows them to scale their use of AI, expanding from individual applications to more comprehensive systems that align treasury processes with organizational goals.
Despite its promise, AI adoption in treasury is not without challenges. Data privacy and cybersecurity top the list of concerns, cited by 57% of finance leaders as major barriers. Treasury teams, responsible for handling sensitive financial data, must prioritize security and establish robust safeguards to protect against breaches.
Integration with existing systems is another obstacle. Legacy infrastructure, common in treasury operations, often lacks compatibility with advanced AI tools, necessitating costly upgrades or custom solutions. Additionally, treasury teams face a skills gap, with limited expertise in configuring and optimizing AI technologies. Building these capabilities requires investment in training or external expertise.
Governance is a growing area of focus. 39% of companies plan to include AI-related risks and controls in financial reporting, while 44% of private companies have procured third-party controls assurance for their AI systems. These measures not only address risks but also build trust in AI’s outputs, ensuring they align with business objectives.
As AI continues to evolve, its potential applications in treasury are expanding. Innovations like Generative AI (GenAI) are opening new possibilities, particularly in dynamic scenario modeling and real-time decision-making. GenAI can simulate complex financial scenarios, providing treasury teams with actionable insights that enable faster and more informed responses to market changes.
The integration of AI is also reshaping treasury’s role within organizations. Traditionally seen as a transactional function, treasury is now becoming a strategic partner, contributing to broader financial and business goals.
AI is no longer a future consideration for treasury. Its ability to enhance efficiency, accuracy, and decision-making is already transforming treasury operations. However, unlocking its full potential requires overcoming barriers such as data security concerns, legacy system limitations, and skill gaps.