Monei, an AI-native financial infrastructure platform that connects payments, banking, investments and insurance through a unified API, has partnered with Scrub.io to integrate real-time fraud monitoring into programmable wallets built for autonomous agents. The companies say the system is operating in live marketplace environments, where AI agents manage procurement tasks and recurring purchases. The partnership links Monei’s transaction orchestration layer with Scrub.io’s behavioural monitoring system. According to the companies, the integration is intended to secure financial activity initiated by software rather than humans. Around the world, payment networks are preparing for similar shifts. Visa has introduced Intelligent Commerce frameworks. Mastercard has announced Agent Pay initiatives. Stripe and PayPal are embedding payments into AI interfaces. Those efforts suggest that payment infrastructure is increasingly being designed for transactions initiated by software. Redesigning payments for software actors For decades, digital payments have assumed a person at the centre of every transaction. Someone reviews a cart, enters credentials, confirms intent and authorises payment. Once authority is delegated to an autonomous agent, that checkpoint disappears, and execution becomes continuous. Mogaji Olarenwaju, founder of Monei, believes that change runs deeper than interface upgrades. “The shift to AI agents transacting autonomously isn't a new feature for the financial system; it's a fundamental redesign of its operating system,” he said. In practical terms, that redesign alters the role of the wallet itself. When software is permitted to spend on behalf of a user or business, the wallet cannot function as a passive store of value. It must actively enforce constraints in real time. “It is not merely a container for funds. It is a policy engine with embedded controls,” Olarenwaju told Condia. Under that model, delegated authority is encoded directly into infrastructure. Spending limits, merchant restrictions and liquidity routing rules are applied automatically, turning the wallet into an execution layer rather than a static account. Rethinking fraud in a machine context The partnership with Scrub.io centres on a problem that becomes unavoidable once machines begin transacting on their own: how do you police abuse when there is no human behaviour to compare against? Most fraud detection systems are calibrated around people. They look for unusual spending patterns, location mismatches or behavioural anomalies that deviate from an individual’s history. Autonomous agents do not behave like people in the first place. Their transaction patterns can be faster, more repetitive and more structured by design. “Human-based fraud systems fail when the ‘user’ is a machine,” Olanrewaju said, noting that the risk model must shift toward continuous monitoring and large-scale pattern analysis rather than static rules. Instead of asking whether behaviour looks suspicious in human terms, the system establishes a baseline for each agent’s delegated authority and flags activity that falls outside those boundaries. The shift moves fraud prevention from post-transaction review to real-time enforcement inside the payment flow. Settlement rails and currency strategy Monei’s architecture connects to existing settlement networks, including card systems and stablecoin rails, while building a separate logic layer intended for continuous machine-to-machine transactions. Olarenwaju argues that the current financial stack was designed for human pacing rather than autonomous execution. “While we connect to existing systems for settlement, the logic, speed, and liquidity for machine-to-machine flows must be optimised from scratch,” he said. That argument arrives at a moment when stablecoins are already reshaping cross-border transfers. Research from the International Monetary Fund shows that stablecoin-based payments can reduce settlement time and lower transaction costs compared with traditional correspondent banking, where funds move across multiple intermediary banks before clearing. In regions where cross-border transfers remain expensive and slow, programmable digital dollars have begun to function as informal settlement infrastructure. USDC, which Monei uses within its stack, has increasingly been adopted for cross-border commerce and treasury management because it allows instant programmable settlement without relying on legacy clearing windows. For autonomous agents executing transactions across time zones, settlement speed is not optional. It is structural. Yet the reliance on dollar-denominated stablecoins introduces a larger question. If AI agents begin executing procurement, subscription payments and trade flows at scale, the currency they default to could shape liquidity distribution across markets. Academic research has warned that widespread stablecoin adoption in emerging economies may weaken local monetary sovereignty if domestic currencies are displaced by dollar-linked digital assets. Olarenwaju frames that risk explicitly. “It will not turn out great for Africa if AI agents make USD their default currency,” he said. “We are actively building a multipolar economy where the Naira plays a major role.” That position aligns with broader efforts across Africa to reduce reliance on the dollar in intra-continental trade. Initiatives such as the Pan-African Payment and Settlement System are attempting to enable cross-border settlement in local currencies rather than routing transactions through offshore dollar clearing systems. The arrival of AI-driven commerce adds another layer to that conversation. If machines begin managing trade flows, the infrastructure they use could reinforce or bypass existing monetary frameworks. In that context, currency choice becomes more than an engineering decision. It becomes a question of how programmable finance will interact with regional liquidity and policy autonomy. Whether autonomous systems accelerate dollar dominance or support local currency rails may depend on how early infrastructure providers design their stacks. From concept to deployment Monei says the infrastructure has moved beyond internal testing. “We are past the experimentation stage,” Olarenwaju told Condia. “We have active partnerships with marketplaces and service providers where AI agents are handling procurement and managing household recurring purchases.” Independent verification of those deployments is limited, and transaction volumes were not disclosed. Still, the claim reflects a broader shift across the payments industry. Infrastructure providers are no longer designing exclusively for human-triggered transactions. They are preparing for systems that initiate financial activity programmatically. If autonomous agents begin managing procurement, subscriptions and investment flows at scale, the focus of payments infrastructure will shift from checkout interfaces to how authority is encoded, enforced and audited. The larger test will not be technological feasibility. It will be trust. Whether businesses and regulators accept software as a direct economic actor may determine how quickly this model moves from edge case to routine practice. For now, the Monei–Scrub partnership illustrates how programmable wallets, continuous monitoring and machine-optimised settlement rails are being assembled into a financial stack built for software actors.
