The CFO’s Treasury Stack Is Learning to Move Money, Not Just Monitor It

Corporate treasury technology has traditionally focused on providing visibility into balances, payments, liquidity, and forecasts. However, as artificial intelligence advances, treasury management systems are evolving beyond mere monitoring to interpreting financial data. Bank of America recently introduced an AI-powered solution called AskGPS, which combines client and account information with AI capabilities to offer treasury insights.

The software can now provide intelligent treasury management reviews, account schematics, fund flow visualizations, and relationship intelligence to identify changing client needs.

American Express recently unveiled a new corporate card and management platform with AI features, while Citi launched a service that enables its bank clients to access multiple cross-border instant payment markets through a single account structure.

As the treasury infrastructure becomes more interconnected through APIs, SWIFT, host-to-host connections, enterprise resource planning (ERP) systems, and real-time payment networks, the integration of AI can help software progress from simply showing where money is to understanding where it needs to be and helping move it there.

The next stage in treasury management involves moving beyond basic observation to execution. Financial software has historically aggregated financial information from various sources; however, the next generation aims to interpret these signals. Once software can understand a company's financial situation, it can raise crucial questions, such as why stop at generating alerts when it can take action based on the insights?

Financial software has spent years compiling information from bank portals, spreadsheets, and enterprise systems. The next generation of treasury software is being asked to interpret this data. While payment execution has become more commoditized, the real value lies in making informed payment decisions. Andrew Ng, Head of Payments and Embedded Finance at Tungsten Automation, emphasizes that companies should focus on an overarching control plane that includes shared data, policy and approval controls, AI-driven recommendations, and access to multiple payment rails.

A PYMNTS Intelligence report found that over 80% of CFOs at large companies are already using or considering adopting AI. The hard problem in treasury management is no longer intelligence but gaining the necessary permissions to act on AI-driven recommendations. As real-time transaction data becomes more prevalent, treasury teams can continuously monitor conditions and recommend or execute appropriate actions, shifting the focus from finance teams periodically determining liquidity needs to software monitoring conditions and providing real-time recommendations.