Riskified Launches Agent Identity Risk Intelligence for AI Agents

What happened

Riskified has launched Agent Identity Risk Intelligence, a new capability designed to help merchants detect and verify when an AI personal assistant — such as Muse — is acting on a shopper's behalf at checkout or during customer service interactions. The e-commerce risk and fraud-prevention provider says the offering is built to identify agentic AI traffic specifically, distinguishing legitimate AI-driven purchases and service requests from fraudulent or impersonated activity.

The move responds to the rapid rise of AI shopping agents and assistants that can browse, select and complete purchases, or handle support queries, without a human directly at the keyboard. Riskified's new tool is positioned as a way for retailers and service providers to establish who — or what — is really behind an interaction before granting trust, access or a transaction.

Why it matters

Agentic commerce is moving from pilot to practice: AI assistants are increasingly able to transact and interact with brands autonomously, and merchants have had few reliable ways to confirm whether traffic claiming to be an AI agent is genuine, spoofed, or malicious automation in disguise. A dedicated identity-risk layer for agents gives businesses a mechanism to extend trust to legitimate AI-driven commerce while keeping fraud controls intact.

For digital transformation and risk leaders, this signals that identity and fraud infrastructure built for human shoppers now needs an equivalent for non-human actors. As more checkout and service volume shifts to agents acting on consumers' behalf, verifying agent identity becomes as foundational as verifying card or account identity has been for the past two decades.

The Renascence take

This launch is less about fraud prevention in isolation and more about the emerging trust architecture of agentic commerce. Every new interaction channel — chat, voice, now autonomous agents — eventually forces a reckoning over who gets authenticated, how, and at what point in the journey.

Most organisations are still designing checkout and service experiences for a single mental model: a human on the other end. Agent Identity Risk Intelligence is an early signal that this assumption is already outdated — and that the next layer of customer experience design isn't just about serving people faster, it's about deciding which non-human actors deserve the same frictionless trust a loyal customer gets. The real behavioral question isn't "is this a bot?" — it's "whose bot is it, and have they earned the same confidence we'd extend to the human it represents?" Operators who treat agent verification as a bolt-on fraud feature will miss the bigger shift: agent identity is becoming a core part of the customer relationship itself.