AI in Treasury: Improving Forecasting, Not Just Automating

A recent study from Stanford Medicine found that while AI-powered chatbots and diagnostic tools can suggest thorough answers and flag critical issues in patient care — such as ensuring tests are ordered before prescribing medications — these digital systems alone aren’t enough. Physicians were able to elicit more complete histories, adapt plans when patients had unusual drug reactions or complex backgrounds, and interpret subtle real-time symptoms that AI couldn’t process. The study’s authors concluded that “human plus computer” consistently outperformed either one alone, demonstrating that technology enhances — but cannot replace — the deep expertise, judgment, and adaptability of skilled medical professionals in complex scenarios.The same can be said for community financial institutions (CFIs) that are integrating artificial intelligence (AI) into their processes. AI is moving rapidly from buzzword to budget line item in treasury departments. In 2025, nearly four in five CFOs are ramping up their AI investments. The draw is greater accuracy, efficiency, and insight in managing liquidity and risk. However, true transformation only happens when technology, data, and human skill are joined together. This is an important lesson as many institutions’ results fall short of their expectations. Why AI for CFIs Today?Why the surge in AI for treasury? The challenges are familiar: volatile interest rates, deposit flight, complex regulatory requirements, and an unrelenting need for real-time reporting. CFIs —especially those with lean teams — see AI as a way to automate repetitive tasks, like downloading bank balances or reconciling exceptions, and to turbocharge analytics for forecasting cash flows, simulating what-if scenarios, and catching fraud before it hits the ledger. While adoption is high, PwC’s 2025 Global Treasury Survey shows that only 26% of firms rate their AI processes as “mature”. Most remain stuck at early stages due to skills gaps or integration bottlenecks. Making AI Deliver on Its PromiseThe institutions making headway are those that view AI as more than just a plug-and-play upgrade. Analysts at J.P. Morgan note that the real value comes when teams couple machine learning with robust domain knowledge. This turns raw data into strategic insight. For example, AI-driven forecasting platforms can run thousands of simulations to project liquidity needs under different scenarios, allowing banks to fine-tune borrowing, lending, and investment decisions. JP Morgan’s research finds that these tools provide “explainable” predictions, giving finance and risk leaders confidence to act, rather than just a “black box” output. Meanwhile, anomaly detection algorithms help identify suspicious transactions early, supporting growing security and compliance demands.Bridging the skill gap is now central. Successful AI programs prioritize upskilling treasury and finance staff, not just buying third-party solutions. This includes collaborative workshops between business users and data scientists, plus transparent performance metrics to help leaders track ROI as projects scale. It’s recommended that banks start with “explainable AI” — platforms where users can audit and adjust recommendations as needed — building trust and buy-in at every level. Strategic Essentials for CFIs

Looking ahead, AI is set to help CFIs become leaner, faster, and more adaptive, but only with the right focus on skills, transparency, and purposeful integration. For those bridging technology and talent, the future of treasury is not just about automated tasks — it is about smarter, more accountable forecasting and risk management.