New safeguards for automated financial deals

The growth of cryptocurrency payments is accelerating, but rising fraud risks from automated transactions—where code performs shopping or payments, have led to calls for stronger protections. On Thursday, Elliptic, a London-based risk-management company, introduced the Elliptic Standard, a framework of eight guidelines to curb fraud in self-executing crypto deals.

Data shows agentic transactions surged by 522% between late May and late August, according to Elliptic’s own research. Though total transaction volumes remain unconfirmed, the firm cited Artemis Analytics figures revealing stablecoin spending reached $33 trillion in 2023. This demonstrates the rapid shift toward automated financial agents in the crypto economy.

Elliptic notes that existing blockchain analysis tools were designed for an era without self-acting agents handling purchases or payments. Blockchain analytics was built for a pre-agentic world, highlighting the need for updated safeguards. The new standard integrates human review and technical controls from the outset, rather than applying them reactively after fraud is detected.

Agentic commerce operates through code acting on users’ behalf, locating products, negotiating terms, and processing payments, but current systems lack inherent fraud prevention. Joanna Marathakis, Circle Internet’s vice president for financial crimes compliance, stressed that oversight must be embedded in design, not bolted on afterward. A criminal only needs to be right once, while businesses face strict compliance obligations to avoid penalties.

The eight rules in the standard address technical and operational protections, including data accuracy, model transparency, and AI security, alongside practical steps like human supervision and staff training. A core principle is business resilience, ensuring fallback rules allow the business to stay in control if automated processes encounter errors. The guidelines also prioritize flexibility, letting firms tailor rules to their unique risks.

Though adoption remains voluntary, the standard’s release reflects industry acknowledgment of the dangers. The balance is clear: automated systems outpace human oversight, but without proper controls, the consequences of failure could outweigh any efficiency benefits.

Simone Maini, Elliptic’s CEO, emphasized the stakes: “We need a new approach … built on an operating system that can execute at agentic speed. A criminal only needs to be right once. Our customers need to be right every time, and that only holds if we stop bad actors at the door, not investigate them afterwards.”