The rise of autonomous AI agents in retail is forcing merchants to overhaul their fraud detection systems to distinguish between malicious bots and legitimate, machine-led customer transactions.

Key Points

• Agentic commerce is projected to account for $190 billion to $385 billion in U.S. ecommerce spending by 2030.
• Online orders driven by large language model referrals increased by more than 1,000% year over year by the end of 2025.
• Current retail fraud models often misclassify authorized AI agents as suspicious traffic, leading to the accidental rejection of legitimate customer orders.
• Retailers must modernize their infrastructure to verify the identity of AI platforms and ensure agents have explicit authorization to act on behalf of human users.
• Businesses are encouraged to provide machine-readable product data, including pricing and shipping rules, to improve discoverability for AI shopping agents.

Why it Matters

Retailers face a significant risk of losing revenue if they fail to adapt their trust models to accommodate this new class of machine-led customers. By modernizing their commerce stacks now, brands can capture emerging demand while preventing the friction that occurs when legitimate agent-driven transactions are incorrectly flagged as fraud.