Highlights
- Ripple Treasury expanded its GSmart AI suite with new tools for forecasting, liquidity, risk, reconciliation and reporting.
- 60% of eligible customers have enabled Risk Insights and 44% have enabled Forecast Insights.
- Every AI-proposed action cites the specific policy clause behind it and requires human sign-off before execution.
- Ripple Treasury processed more than $13 trillion in customer transactions in 2025.
Ripple's enterprise treasury division has rolled out a major expansion of GSmart, its policy-governed AI system, adding new capabilities spanning cash forecasting, liquidity stress-testing, risk simulation, reconciliation and reporting. The unit, built on the former GTreasury platform that Ripple folded into its business, says the goal is to let finance teams move faster on routine treasury decisions without giving up the audit trail and human sign-off that corporate controllers require.
The new toolset includes SmartLedger for real-time cash-position visibility tied directly into a client's existing accounting and ERP systems, Forecast Insights for spotting patterns and variances between projected and actual cash flows, Liquidity Scenarios for running multi-outcome stress tests, and a Risk Management layer that runs CFAR and Monte Carlo simulations against a company's exposures. According to Ripple Treasury's own announcement, the system already shows meaningful uptake: 60% of eligible customers have turned on Risk Insights to flag exposure anomalies and policy breaches, while 44% are using Forecast Insights to catch emerging liquidity gaps before they become problems.
Governance Built Around the Approval Button
What sets GSmart apart from a generic AI copilot is how narrowly Ripple has scoped its autonomy. Each agent monitors its assigned process, proposes a specific action, and cites the exact internal policy clause that justifies it — but nothing executes until a human approves it. Ripple Treasury CPO Mark Johnson summarized the design philosophy directly: “Every recommendation is explainable, every insight is auditable, and every action is rooted in your own data.” The system also does not train its underlying models on client data, running in inference-only mode, and the company says it aligns with ISO/IEC 42001, ISO/IEC 27001, GDPR and the EU AI Act — a compliance list aimed squarely at the corporate treasurers and CFOs who are GSmart's actual buyers, not retail crypto users.
Why This Matters Beyond Treasury Desks
The expansion lands at a moment when Ripple has been steadily repositioning itself away from a pure XRP story and toward enterprise financial infrastructure, layering custody, stablecoin settlement and now AI-driven treasury management on top of its blockchain payments business. Ripple Treasury's scale underscores how far that pivot has gone: the unit says it processed more than $13 trillion in transactions across its customer base in 2025, a volume figure that dwarfs most crypto-native platforms and puts it in direct competition with legacy treasury-management vendors like Kyriba and SAP rather than with other blockchain firms. It also fits a broader pattern of finance-adjacent companies wiring AI agents into money-moving workflows under tight guardrails, echoing similar governance-first AI rollouts elsewhere in the sector as institutions try to capture AI's efficiency gains without losing control over who — or what — can authorize a transaction.
What to Watch Next
The figures Ripple has disclosed so far measure feature activation, not outcomes, so the next real signal will be whether adoption of Risk Insights and Forecast Insights keeps climbing among the roughly 40-56% of eligible customers who haven't yet turned them on. Longer term, the more consequential question is whether Ripple links GSmart's liquidity and forecasting tools directly into its RLUSD stablecoin and XRP Ledger settlement rails, which would turn this from a standalone treasury product into a deeper wedge for pulling corporate finance workflows onto Ripple's own blockchain infrastructure.
Related: 17B Jeonbuk Bank Taps Ripple to Replace SWIFT for Cross-Border Payments
FAQ
What is GSmart?
GSmart is the AI system built into Ripple Treasury's platform (formerly GTreasury) that helps corporate finance teams with cash forecasting, liquidity stress-testing, risk simulation and reconciliation.
Does GSmart execute financial transactions automatically?
No. Each AI agent proposes an action and cites the policy clause behind it, but a human must approve it before anything is executed.
How many companies use Ripple Treasury?
Ripple Treasury says it serves more than 1,000 customers who processed over $13 trillion in transactions in 2025.
How widely have customers adopted the new AI features?
60% of eligible customers have enabled Risk Insights and 44% have enabled Forecast Insights, according to Ripple Treasury's announcement.
