The Holiday Fraud Test Is Weeks Away and the Agent Stack Isn’t Ready

With the peak shopping weekend beginning November 26 and running through Cyber Monday on November 30, merchants have approximately five weeks from October 20 to finalize their infrastructure. The industry spent October deploying a new defensive stack – identity, fraud, checkout, and insurance layers – yet none of it has faced the volume of holiday-scale traffic. How these systems hold under peak load is the open question.

The October Sprint

The industry spent October building a defensive perimeter around agentic commerce. Between late September and mid-October, a series of infrastructure launches arrived: Mastercard’s trust services, Cequence’s identity layer, Forter’s orchestration engine, and Harbor’s insurance seed. PayPal and Glance introduced agentic checkout capabilities on October 15. These tools are arriving just as the calendar turns toward the high-volume period of Thanksgiving and Cyber Monday.

Activity levels are rising. According to a vendor-reported release from Forter, agentic orders across its network have increased more than fourfold since August. Automated fraud has doubled in the same two-month window. The figures are scoped to Forter’s own network and carry no published dollar baselines. The pressure on merchants is two-sided: a new revenue channel on one hand, and, in Forter’s words, bad actors “attempting to blend in to this traffic by using malicious agents or taking over legitimate ones to scale fraud and abuse” on the other.

The Trust Gap

Market readiness is varied. Cisco research from March 2026 found that 85% of organizations are experimenting with AI agents, but only 5% have moved them into broad production. Consumer behavior mirrors the gap. Per the Marketplace Shopping Behavior Report 2026, 58% of shoppers use AI for research, but only 17% feel comfortable completing a purchase through an agent; in the trust funnel analysis, 37% start a purchase journey through an assistant and 76% demand a human review step before money moves.

What Breaks First

With the peak shopping weekend opening November 26, the industry is running a live, un-tested experiment. On the current deployment timelines and market data, four points look most vulnerable:

• Fraud decisioning. Automated fraud doubled alongside the fourfold rise in agentic orders, and malicious agents blend into legitimate traffic. Distinguishing a trusted agent from a fraudulent one at peak volume is the first point of failure.
• Merchant integration. Forter’s orchestration layer is only “available now for interested Forter customers,” integration windows are closing, and fewer than 11% of small merchants are ready for AI shopping agents.
• The consumer approval gate. Because 76% of shoppers require human intervention, the efficiency agentic commerce promises may be offset by a surge in support and returns volume during the busiest days of the year.
• Liability exposure. The financial responsibility for agent-driven errors remains largely unassigned. 93% of merchants believe the AI provider should bear the loss for incorrect purchases, and traditional carriers are introducing AI exclusion endorsements – leaving current insurance capacity, often capped at $15-25 million per organization, a fraction of potential multi-billion dollar exposure, as the liability wall keeps score.

As Eran Vanounou, Forter’s chief technology and product officer, put it: “Every merchant is now effectively opted in to agentic commerce, whether or not they’ve prepared for it. The choice merchants have isn’t whether agentic traffic engages with their site this holiday season; it’s whether they are equipped to control, optimize and drive revenue from it.” The challenge for the next five weeks is an economic test: whether this stack can protect margins when automated traffic – and the risk riding with it – reaches its annual peak.