How AI and Stablecoins Are Reshaping Corporate Treasury
1. How is AI transforming traditional treasury functions such as cash forecasting, risk management, and liquidity planning?
AI moves the treasury function from reactive to predictive. Cash forecasting historically leans on historical averages and a treasurer’s instinct, and analysts at most enterprises can spend 30-40% of their time generating forecasts and chasing variances by hand. But implementing the right AI-powered treasury technology strategy can handle extraction, pattern detection, and narrative generation, which converts all that time into forward-looking visibility. We typically see a 30% increase in forecast accuracy within the first full reporting cycle after go-live.
Risk management is also evolving quickly, as machine learning pinpoints exposures that a human analyst would struggle to catch in time. Liquidity planning is no longer about end-of-day snapshots, either. Corporate treasurers can see where cash will be hours from now across every entity and currency, including how much is sitting idle, which opens the door to generating yields through digital currencies intra-day, overnight, and on weekends. That kind of structural visibility is something the treasury function has needed for years.
2. What are the biggest challenges organizations face when integrating AI into treasury operations?
One of the biggest barriers is something most vendors will not say out loud, which is operational toil. AI adoption is slowed down by the operational backlog teams already carry. When your people are buried in manual cash positioning, spreadsheet forecasts, and reconciliation exceptions, there is no capacity left to stand up something new. That is usually the real drag, not some abstract data quality problem.
Governance is the next piece, and I strongly argue it is a design requirement rather than a barrier. Every AI output needs to be traceable and explainable to the CFO and audit committee, and that has to be built in from day one, ideally, to avoid a painful retrofit later.
Security is another concern that warrants understanding. Treasurers need confidence that their data is not being shared with third parties or used to train models outside their environment. That means inference-only processing, regional data residency, and a complete audit trail on every AI interaction. If your vendor cannot tell you exactly where your data is processed and guarantee it is never retained or used to train models, you have not solved the security question.
3. How can AI-driven insights improve real-time decision-making for corporate treasurers?
Reliable real-time decision-making depends on having a crystal clear picture of cash, exposures, and obligations at any given moment, and AI is what makes that picture actionable. Instead of waiting on a morning report, you can see anomalies flagged as they occur and get recommended actions aligned to current market conditions. For example, suppose you are evaluating whether to draw on a credit line, move funds between subsidiaries, or hedge an exposure. You can now do that based on models that have already weighed dozens of variables and can show you their reasoning.
The next stage, which is closer than most treasurers realize, is genuinely agentic. The legacy model for liquidity is a static worksheet that calculates each account’s funding gap independently, with no awareness of what is happening across the rest of the entity structure. The next generation is an agent that evaluates total demand against total supply across every account simultaneously, reasons about which funding source is optimal under constraints, and generates a coordinated set of proposals with plain-language rationale before the treasurer opens the system. The treasurer reviews and approves. The analytical reconstruction that used to take hours happens before they arrive. That is what real-time decision-making actually looks like in treasury not faster dashboards, but intelligence that precedes interaction.
4. In what ways is AI helping treasurers enhance fraud detection and financial security?
Payment fraud is increasingly sophisticated with social engineering, deepfakes, and compromised vendor accounts impersonating trusted parties. Rules-based systems were not built for these kinds of threats. AI counters it by establishing behavioral baselines across every payment flow and flagging deviations a static system would miss. An unusual beneficiary, a payment that breaks from historical timing patterns, a request arriving through an atypical channel…any of these can trigger review before funds leave the building. Machine learning models also sharpen over time, calibrating against both confirmed fraud and false positives. The key is that the system explains why it flagged something, not just that it did. A reviewer who sees a flag with no context is no better off than one who missed it entirely. Explainability is what turns detection into prevention.
5. What role do stablecoins play in modern treasury management, especially for cross-border payments and liquidity optimization?
The use cases that are becoming real right now are not speculative, and intracompany transfers are at the top of the list. Moving value between subsidiaries across borders using digital rails greatly reduces the FX conversion costs and settlement delays that accumulate across a global enterprise.
Idle cash is the other big story here. Balances that historically sat dormant from Friday afternoon to Monday morning earning nothing can now generate intraday, overnight, and weekend yield inside a compliant enterprise treasury system. That changes how a global treasury can operate, because on-chain networks do not observe banking holidays or cutoff times. I am seeing real adoption among companies that have outgrown what traditional cross-border infrastructure was built to do.
6. How do stablecoins compare to traditional banking systems in terms of efficiency, cost, and risk?
On efficiency, digital rails are very hard to beat. Settlement is near-instant, available continuously, and not dependent on a chain of intermediary banks each taking a fee and adding latency. Costs also tend to be lower, particularly on corridors where correspondent banking is expensive or unreliable. Risk is where I think the conversation gets more nuanced. Treasurers need to evaluate issuer solvency, reserve transparency, custody, and smart contract integrity with the same thoroughness that they would apply to any counterparty.
Treasurers also need to understand that not all digital enablement claims are equal. Some vendors offer stablecoin capability that still routes through the same legacy correspondent banking infrastructure underneath. That is a digital label on an old problem, and it does not deliver the efficiency or cost benefits treasurers are actually looking for.
7. What regulatory and compliance challenges should enterprises consider when adopting stablecoins in treasury operations?
Regulatory pictures vary by jurisdiction, and any treasurer operating globally needs to track developments in each market where they hold or move stablecoin balances, for example with the GENIUS Act in the U.S. and MiCA in the EU. Accounting treatment is another open area. There are important questions around classification, impairment, and disclosure that finance teams need to work through with their auditors. Anti-money laundering and sanctions compliance need careful thought too, which is why stablecoins like Ripple USD (RLUSD) that are subject to regulatory oversight under a NYDFS Trust Charter, with custody providers like BNY Mellon are considered the gold standard for institutional use cases.
What I tell treasurers and CFOs is that none of this should discourage adoption, but it does mean the audit trail requirements and compliance complexity are not an IT problem. They are a treasury problem, and they need to be solved inside a system that was built for enterprise governance from the start rather than bolted on after the fact.
8. How do you see the future of treasury evolving with the combined adoption of AI and stablecoin-based financial systems?
The technologies reinforce each other in a way that fundamentally changes what corporate treasury can do. AI gives treasurers the intelligence to make faster decisions, and digital rails give them the infrastructure to act on those decisions without waiting on legacy settlement. Put them together and you get a treasury function that operates continuously, allocates capital with much greater precision, and responds to market conditions in something close to real time.
The treasurer of the next decade will spend less time chasing balances and reconciling statements, and spend more time on capital structure, counterparty strategy, and how to deploy liquidity for competitive advantage. That shift is already underway at the most forward-looking enterprises, and the gap between early movers and everyone else is compounding every quarter.
About Mark Johnson:
Mark Johnson is VP, Global Product, at Ripple Treasury. He was previously the Chief Product Officer at GTreasury; Ripple acquired the treasury management software company in 2025. Mark has 15+ years of expertise in FinTech, Payments, and SaaS. He has also held product leadership roles at BigTime Software, Deluxe, and The Home Depot.
