Treasury AI adoption: drive the change or chase it

There is a quiet revolution happening inside the world's most forward-looking finance organizations, and most treasury teams are watching it from the hallway. Artificial intelligence is no longer a pilot program, a buzzword in a vendor deck, or a problem for IT to solve. Treasury AI adoption is actively reshaping how cash is managed, how risk is quantified, and how financial leadership earns its seat at the strategic table.

The question is no longer whether AI will transform treasury. It already is. According to Kyriba's 2026 CFO Survey, approximately 92% of global CFOs are integrating AI into some of their processes and decision-making. The only question is whether yours will be among those driving that transformation, or scrambling to catch up with those who did.

The real risk is inaction

Let's be honest about what the hesitation often looks like: concern about data security, skepticism about ROI, uncertainty about where to start, and (perhaps most dangerously) the comfort of the familiar. These are understandable instincts. Treasury has always been a function defined by precision and caution. But those same instincts, left unchecked, become an existential liability when the competitive landscape is moving at the speed of AI adoption.

Consider what's at stake. While your team manually reconciles bank statements and consolidates cash positions across 40 entities, an AI-enabled competitor is receiving real-time global liquidity intelligence before the market opens. While your analysts are building this week's FX exposure model in Excel, a peer institution's system has already flagged a correlated risk in a currency pair you haven't touched yet. The gap isn't just operational efficiency; it is strategic foresight. And in treasury, the cost of delayed foresight is measured in basis points, counterparty risk, and missed working capital optimization, sometimes in the tens of millions.

Waiting for AI to be "proven" in treasury is like waiting for the internet to be proven in banking. The proof is already in, and the cost is being paid by those still waiting.

What AI actually delivers: cash flow forecasting, FX risk, and beyond

Strip away the hype and the financial case for treasury AI adoption is concrete and compounding. Cash flow forecasting, historically the most labor-intensive and least accurate function in the treasury toolkit, is being transformed by machine learning models that learn from ERP data, payment patterns, and external signals simultaneously. According to Kyriba customer data, organizations deploying AI forecasting are reporting 30-50% improvements in forecast accuracy, translating directly into lower precautionary cash buffers and higher yield on deployed liquidity.

On the risk side, AI is enabling dynamic FX hedging programs that adjust in near-real-time to exposure changes rather than quarterly rebalancing cycles, a structural advantage in volatile macro environments. For organizations with complex intercompany structures, AI-powered netting and pooling optimization is consistently surfacing working capital improvements that manual treasury operations simply cannot detect at the required speed or granularity.

Treasury teams deploying AI automation in cash positioning, payment processing, and reporting are reclaiming 15-25 hours per analyst per week, time that elite treasury organizations are reinvesting into capital structure strategy, M&A support, and board-level financial risk advisory. That is not incremental improvement. That is a fundamental repositioning of what treasury contributes to the enterprise.

Fear is the most expensive line item on your balance sheet

The AI fears circulating in treasury circles deserve acknowledgment, but not accommodation. Concerns about model explainability are legitimate; the answer is to demand transparency from vendors and build internal AI literacy, not to abstain. Concerns about data security are valid; the answer is rigorous governance frameworks, not a blanket moratorium on adoption. Concerns about job displacement deserve a thoughtful response: treasury teams that adopt AI don't shrink; they evolve. The analysts who once built cash reports become the strategists who interpret AI-generated intelligence for the CFO and the board.

What is worth examining honestly is whether vague discomfort is masquerading as prudent risk management. Every month a treasury organization delays AI adoption is a month of compounding disadvantage: in forecast quality, in working capital efficiency, in FX risk management, and ultimately in the credibility of finance leadership as a strategic partner to the business.

AI for CFOs and treasurers: lead from the front, or explain why you didn't

CFOs and treasurers are uniquely positioned to lead enterprise AI adoption, not just within finance, but as a model for the broader organization. Treasury sits at the intersection of data, risk, and strategy. The function already commands the systems, the governance instincts, and the cross-functional relationships needed to deploy AI responsibly and at scale. The organizations that seize this moment will not simply become more efficient treasury departments. They will become the intelligence engines of their enterprises, providing the real-time financial visibility and predictive risk insight that transforms how the C-suite makes decisions.

That is a future worth leaning into. The technology is mature enough to deliver. The business case is clear enough to defend. The only variable that remains is leadership conviction.

Imagine a treasury team that had fully embraced AI in cash forecasting, FX risk, liquidity optimization, and reporting. How much more would they know? How much faster would they move? And how much more would the business trust them with? If the honest answer unsettles you even slightly... what, specifically, is holding you back?