Riskified's Agent Identity Risk Intelligence Answers the Wrong Question First
On October 8, 2026, Riskified (NYSE: RSKD) launched Agent Identity Risk Intelligence — a system that identifies the AI personal assistant behind an order or customer-service request (Meta's Muse, Instinct, dots in ChatGPT), resolves it to the real consumer, and returns a real-time risk signal to the merchant. First platform: Shopify. Delivery: Riskified's existing APIs, the AI Agent Approve MCP…
On October 8, 2026, Riskified introduced Agent Identity Risk Intelligence, a system designed to identify the AI personal assistant behind an order or customer-service request, trace it back to the actual consumer, and provide merchants with a real-time risk signal. The platform initially launched with Shopify integration and will be available through Riskified's existing APIs, AI Agent Approve MCP server on AWS Marketplace, Decision Studio, and Zendesk.
The company plans to launch the feature to customers in limited beta by December 2026, with a generally available release expected by 2027.
Riskified highlighted the distinction between authenticated and trusted agents, stating that an agent authorized by a stolen account or a refund-as-a-service ring is still considered fraud, regardless of the machine-speed nature of the fraud. They provided examples of live tests on a decision gate, demonstrating how fraud rings could bypass identity checks, while a legitimate shopper with a single refund was still flagged for review.
The company explained that Identity risk and instruction judgment are two separate instruments, emphasizing the need for both to be integrated into a merchant's execution path. They shared a link to a dated table and full receipts, providing an exact curl to score your own agent's instruction and a table to assess the risk.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.