Bank of America wants AI to do treasury's grunt work

Treasury teams have always had plenty of data. The problem is sorting the useful stuff from the noise. Bank of America thinks AI should do more of that heavy lifting, and this morning it added a new Payments Insights feature to CashPro to help do it.

The tool looks at payment efficiency, cross-border flows and working-capital performance, then compares a company’s numbers with peers in its industry. BofA says the point is simple: spend less time gathering and cleaning data, and more time deciding what to do next. It’s part of CashPro Data Intelligence, the bank’s AI-powered treasury analytics suite, which already includes cash forecasting and fraud-security scoring.

CashPro handled 213 million payments in the first half of 2026, up 10% from a year earlier. Jennifer Sanctis, who runs the product for Bank of America, said companies are not tearing up their treasury org charts. They’re trying to free people from manual work so they can focus on analysis and decisions that actually move the business.

That shift is showing up in the skills treasury teams need. Sanctis said staff now have to bridge finance, technology and data, while companies keep pushing on straight-through processing and cost control. Deloitte is seeing the same direction of travel, saying agentic AI is moving beyond pilots and into workflows that cover forecasting, risk analysis and fraud detection directly inside payment systems.

The practical version is less glamorous than the hype. Instead of an analyst dragging payment data into a spreadsheet and hunting for patterns, the system surfaces trends, opportunities and risks through dashboards and peer comparisons. Laura Fox, treasury director at Allegis Group and a member of BofA’s CashPro Boards, said the real value is not having more data. It’s understanding it fast enough to act on it. And that may be the real test for treasury teams now: not whether AI shows up, but whether people are being trained quickly enough to use it well.