Ripple has expanded GSmart, the artificial-intelligence capability within Ripple Treasury, with orchestrated agents designed for forecasting, liquidity management, risk, reconciliation and reporting. The company announced the expansion on September 10 and described GSmart as the industry’s first governed AI platform for enterprise treasury, although that claim has not been independently established.

GSmart is not a new product. GTreasury introduced the platform in June 2025, before Ripple acquired the company. The latest release adds a broader catalogue of task-specific agents, a policy layer called Knowledge Studio and Ask GSmart, a conversational interface for treasury data and reporting. Ripple’s official announcement provides further details.

The main change is that GSmart now connects separate AI capabilities through a common policy and approval layer. Instead of treating forecasting, risk and reporting as isolated tools, Ripple is placing them within the workflows treasury teams already use to review cash, liquidity and financial exposures.

Human approval remains part of the process

GSmart separates treasury tasks that require mathematical certainty from areas where AI can help interpret information. Deterministic software performs financial calculations, while GSmart interprets policies, identifies patterns and explains proposed actions. Companies can use Knowledge Studio to define the controls against which those proposals are evaluated.

For example, a liquidity agent could identify an imbalance between accounts, while a risk tool could flag an exposure anomaly or a policy breach. The system is designed to cite the rule supporting its recommendation, leaving treasury personnel to decide whether the proposed action should proceed.

That distinction is important because Ripple’s announcement does not describe a system that freely reallocates corporate cash or trades digital assets. Instead, it describes a recommendation system intended to operate within the approval and audit requirements already used by finance teams.

GTreasury provides the enterprise distribution

Ripple’s $1 billion acquisition of GTreasury, announced in October 2025, provided the distribution base for the strategy. The transaction gave Ripple a treasury platform serving more than 1,000 customers across 160 countries, along with more than four decades of experience in corporate treasury operations.

GTreasury already connects treasury and finance teams with banks and enterprise resource-planning systems. It gives customers a consolidated view of the data used for payments, forecasting and risk management. Those connections to existing financial processes create an established environment in which an AI layer can operate, rather than requiring customers to use a separate chatbot.

The value of the acquisition therefore extends beyond the software itself. If GSmart becomes part of the routine process customers use to assess cash positions and liquidity requirements, Ripple would be operating closer to the point at which financial decisions are made.

Ripple is building potential execution routes

Ripple Treasury has also been developing infrastructure beneath that decision layer. In April, the company launched capabilities that allow customers to view and manage bank cash, XRP, RLUSD and assets held through external custodians in one system.

Coindoo previously reported on Ripple’s unified cash and digital-asset treasury platform, which is intended to bring on-chain and traditional balances into the same treasury environment. Ripple says its treasury platform facilitated $13 trillion in customer payment volume during 2025.

That figure does not indicate how much money GSmart influences or manages. It does, however, show the scale of the existing system to which Ripple is adding AI capabilities. Approved recommendations could eventually connect with Ripple’s payment, custody and digital-asset services where those products meet a customer’s needs and are available.

This does not mean every GSmart user will use stablecoins or XRP. The more immediate point is that a recommendation concerning liquidity, payments or idle cash can be considered alongside both traditional and digital-asset balances.

Governance is a barrier to wider agent adoption

Ripple’s focus on controls reflects a broader challenge facing enterprise AI. In a survey of 360 IT application leaders, Gartner found that only 15% were considering, piloting or deploying fully autonomous AI agents. Only 19% said they had high or complete trust in vendors’ protections against hallucinations, while 74% viewed agents as a new attack vector. Gartner published the findings in its September 2025 survey release.

Those figures help explain why Ripple is emphasizing traceable recommendations and human sign-off instead of maximum autonomy. In treasury operations, a poor recommendation can affect payment timing, liquidity, compliance and financial reporting. As a result, the source of a proposed action can matter as much as the action itself.

The controls are only as reliable as the customer’s underlying data and policy settings. A cited rule can make a recommendation easier to review, but it cannot correct inaccurate inputs or poorly configured limits.

Adoption figures do not establish financial impact

Ripple says that 60% of eligible customers have enabled Risk Insights and 44% are leveraging Forecast Insights. These figures indicate that users are activating parts of GSmart, but they do not show how many customers were eligible, how often the tools are used or how frequently recommendations lead to approved action.

The company has not provided results concerning time saved, forecast accuracy, error reduction, financial returns or the value of funds influenced by GSmart. Without those measures, feature enablement does not demonstrate that the AI system has improved treasury performance.

The next test is whether recommendations lead to action

Ripple has built a system intended to place AI within existing corporate treasury controls rather than outside them. Its $1 billion GTreasury acquisition gave the company the customer base and financial workflows needed to pursue that strategy, while GSmart serves as the decision layer being added on top.

The next proof point will be whether treasury teams accept the recommendations, how often approvals result in action and whether customers achieve measurable improvements in forecasting, liquidity management, risk monitoring or operational speed.

This article is for informational purposes only and does not constitute financial, investment or treasury-management advice.