Cash management moves up the agenda
In its guidance on cash forecasting, U.S. Bank argues that manual processes and models focused on past performance can leave companies poorly equipped for economic uncertainty and rapid market changes. Inaccurate forecasts can leave capital tied up when a business could otherwise use it for growth.
AI cash flow forecasting combines financial data with predictive analytics to estimate incoming payments, outgoing payments and future balances. The bank says these estimates can be more accurate than traditional forecasts, giving executives a stronger basis for assessing liquidity and allocating capital.
The U.S. Bank Fall 2026 CFO Insights Report found that 32% of CFOs placed improving cash flow among their three leading finance priorities for the next 12 months. Deploying technology within finance made the top three for 28%.
Protiviti's 2026 Global Finance Trends Survey also highlighted pressure on cash management: 83% of CFOs ranked it among the three areas needing the most attention because of economic and trade policy volatility. The greatest emphasis came from financial services at 61%, manufacturing and distribution at 55%, healthcare at 45%, and consumer products at 41%.
Adoption grows, with reported forecasting gains
Financial forecasting was the leading AI application in Protiviti's survey, published Aug. 26, 2026. Overall, 77% of finance organizations used AI, while adoption specifically for financial forecasting increased from 58% to 76% year over year.
KPMG's Global AI in Finance 2026 report found that active AI use in finance had risen from 30% in 2024 to 75%. Its March 2026 survey covered 1,013 senior leaders in 20 countries. Among respondents, 64% reported substantial gains in forecast accuracy, and 71% said AI's return on investment in their finance function met or surpassed expectations.
American Express's 2026 CFO Survey, covering 999 senior finance executives in 14 countries, found that 59% said AI was already producing tangible cash flow improvements. Improving cash flow forecasting was an intention for 64% of respondents.
Reliable data comes before useful forecasts
Dominic Venturo, U.S. Bank's senior executive vice president and chief AI, data, and innovation officer, identifies fragmented data as a central obstacle. Receivables and payables ledgers, enterprise resource planning systems, and bank accounts across subsidiaries and business units all hold information needed for cash management. Companies need a dependable, repeatable way to bring that information together, including accurate current balances.
Companies can connect their banking and financial systems through APIs or establish continuously updated data feeds. In the American Express survey, 50% of executives said they used real-time transaction data to improve cash flow forecasting and working capital management.
With those connections in place, AI tools can analyze transaction patterns and movements through accounts receivable and payable. They can also flag unusual activity that may indicate supplier overcharging or fraud. U.S. Bank says real-time AI systems can help companies intervene before such problems cause financial losses. Businesses may develop these tools internally or obtain them from banks and other technology partners.
Dashboards make the resulting information available according to each user's role. Transaction staff may need operational controls, while executives need forecasts and a companywide view. Senior finance staff can track exceptions and anomalies and oversee transactions that exceed policy limits. Companies can build dashboards using internal and bank data or use bank-provided tools and third-party systems.
Applying forecasts to payment decisions
One use for better cash visibility is deciding whether to take an early-payment discount. A supplier might allow 30 days to pay but offer a lower invoice amount for payment within 10 days. Accepting requires a comparison between the discount and the cost of using cash sooner.
Venturo says the calculation becomes much harder when it extends across hundreds of millions or billions of dollars in payments, especially when decisions must be made in real time. Forecasting tools can bring together current balances, available liquidity, and the costs of cash and credit. They can then estimate future balances, assess the marginal cost of a decision and recommend a course of action.
The inputs also change. A large project payment can raise balances, a major outlay can draw them down, and borrowing costs can shift. U.S. Bank argues that teams relying on incomplete information and manual workflows cannot capture the full value of automating this analysis.
Payment methods offer another opportunity. Some channels cost less than others; corporate credit cards may provide rebates, while other options may secure better supplier terms. Finance teams need to assess those benefits alongside their effects on future cash and liquidity. The same considerations apply to accounts payable decisions about discounts and accepted payment methods.
The bank says such analysis could help finance reduce costs, increase revenue and use capital more efficiently. Venturo argues that the underlying data also allows treasury teams to document and audit the value they create, supporting a role as a strategic partner rather than solely a cost center.
Human oversight and documentation remain essential
U.S. Bank presents predictive analytics as support for finance and treasury professionals, not a replacement for them. Venturo's distinction is between rapidly interpreting information scattered across systems and deciding what action to take. That second task still requires human judgment, and finance staff need to remain vigilant as AI tools improve.
Protiviti's findings show limits in how organizations manage their use of AI. Only 35% described themselves as highly or moderately effective at measuring its return on investment, and just 14% were using it under a defined strategy.
BDO's 2026 Audit Innovation Survey examined AI's role in financial estimates and judgments subject to external audit. Conducted Aug. 6-28, 2026, it surveyed 200 senior finance leaders at U.S. public and private companies with revenue ranging from $250 million to $10 billion.
Among those organizations, 99% used AI to produce or inform at least one estimate or judgment reviewed by their external auditor. Cash flow or earnings forecasts were the most common application, cited by 62%. Yet only 49% said their documentation and processes were sufficient to substantiate every AI-assisted output.
