Ripple Brings AI Into Corporate Treasury
Ripple is expanding its GSmart artificial intelligence platform with a new set of capabilities designed to help corporate treasury teams manage forecasting, liquidity, risk, reconciliation and reporting. The company announced the expansion on September 10, positioning GSmart as an AI system built directly into treasury workflows rather than a standalone chatbot.
The announcement comes as companies experiment with increasingly autonomous AI agents. Gartner estimates that an average Fortune 500 enterprise could have more than 150,000 AI agents in use by 2028, compared with fewer than 15 in 2025. Yet only 13% of organizations surveyed by Gartner said they believe they have appropriate AI-agent governance in place.
That governance gap is particularly important in treasury departments. Treasury teams deal with cash, liquidity, financial risk and payments, meaning an incorrect automated decision can have direct financial consequences. Ripple’s approach is therefore built around keeping humans responsible for consequential actions rather than allowing an AI system to operate without controls.
GSmart separates financial calculations from AI interpretation. According to Ripple, deterministic financial engines handle the calculations, while AI interprets company policies, identifies patterns and explains recommendations. A treasury professional then retains approval authority before a financial action is executed.
The platform’s agent system covers several areas of treasury management. Agents can monitor forecasting and planning, liquidity, risk, reconciliation and reporting processes, then propose specific actions. Ripple says each recommendation can reference the relevant policy clause before being sent to a human for approval.
Governance Becomes the Main AI Feature
One of the more important components is Knowledge Studio, which acts as the policy and governance layer for GSmart. Treasury teams can define organizational rules and controls that determine how the AI operates. Proposed actions are checked against those controls before being escalated to a person.
Ripple is also adding Analytics Studio and Ask GSmart. The conversational system allows treasury teams to query their financial data and obtain insights without manually working through multiple reports. This is aimed at reducing the time needed to identify liquidity gaps, unusual exposures and other potential issues.
The company says GSmart is already being used by its enterprise customer base. According to Ripple, 60% of eligible customers have enabled Risk Insights, which identifies exposure anomalies and policy breaches, while 44% are using Forecast Insights to compare projected and actual cash flows and identify emerging liquidity problems.
Those adoption figures are notable because they show the technology is not being presented solely as a future product. Ripple says GSmart is already in production across its enterprise customer base, with the latest release expanding the number of treasury functions where the system can provide AI-assisted analysis and recommendations.
Security and auditability are also central to the product’s positioning. Ripple Treasury says every GSmart AI interaction is logged with a unique trace ID, giving organizations an audit trail of how the system was used. The company also emphasizes explainability rather than treating the AI as an opaque decision-making system.
That approach reflects a broader problem facing enterprise AI. Gartner has warned that uncontrolled agent proliferation can create security, compliance and operational risks. Its research recommends clear policies, agent inventories, defined permissions and ongoing monitoring rather than simply allowing organizations to deploy autonomous systems without restrictions.
Related: Ripple CTO Emeritus David Schwartz Says XRP Could Surpass Bitcoin With One Major Catch
Ripple’s strategy is to make those controls part of the treasury platform itself. That could become increasingly important as finance departments begin using AI for decisions involving cash positions, liquidity and risk. The question for CFOs is no longer simply whether AI can produce an answer, but whether the organization can explain why that answer was produced and who approved the resulting action.
The expansion also connects GSmart with Ripple’s broader push into digital-asset treasury management. Ripple says its treasury platform is designed to give companies visibility across traditional cash and digital assets while supporting payments, liquidity management, forecasting and risk management from a single system.
For Ripple, the bigger opportunity is therefore not simply adding another AI assistant to corporate software. It is positioning AI as an operating layer for treasury while keeping financial controls, audit trails and human approval inside the same workflow.
If enterprise AI agents become as widespread as Gartner expects, governed execution could become as important as intelligence itself. Ripple’s latest GSmart expansion is built around that idea: let AI find patterns, explain risks and recommend actions, but keep the final financial decision in the hands of the people responsible for the company’s money.
About The Author
