Wednesday, March 04, 2026
Riskified, a global leader in ecommerce fraud prevention and risk intelligence, has announced the expansion of its AI Agent Intelligence platform to help merchants protect their emerging AI-powered shopping assistants. As retailers increasingly deploy conversational AI tools on their websites to enhance customer engagement and personalization, Riskified aims to ensure that these new digital interactions remain secure from fraud and abuse.
The ecommerce industry is rapidly embracing artificial intelligence to transform how customers shop online. Many retailers are introducing AI-driven shopping assistants that provide real-time product recommendations, personalized offers, and seamless customer service experiences. According to research from McKinsey & Company, this trend is accelerating, with approximately 82% of retail organizations already running generative AI pilot programs aimed at reinventing customer service.
However, while these AI assistants create new opportunities for better customer engagement, they also introduce potential risks. Fraudsters are becoming increasingly sophisticated and may attempt to exploit automated shopping systems for fraudulent transactions, refund abuse, or promotional manipulation. Recognizing these challenges, Riskified is expanding its platform to provide a robust security layer for merchant-driven AI agents.
As retailers develop conversational AI assistants capable of offering personalized recommendations and loyalty-based incentives, Riskified contributes an additional intelligence layer designed to keep these interactions safe. The company analyzes global ecommerce purchase data across a broad network of merchants, enabling it to detect patterns and risk indicators that individual retailers may not be able to identify on their own.
“Merchants looking to launch their own virtual shopping assistants have a home-field advantage maintaining direct, personalized relationships with their shoppers,” said Assaf Feldman, CTO and Cofounder of Riskified. “Riskified’s role is to serve as the definitive risk intelligence layer to both enhance and secure AI agent interactions. By augmenting a merchant's proprietary customer history with insights from our vast multi-merchant network, we help to ensure that every transaction and claim enacted through a merchant’s AI agent is a verified identity.”
To strengthen protection for AI-powered retail interactions, Riskified introduced several new capabilities tailored specifically for conversational shopping assistants.
One of the key additions is AI Agent Identity Signals, which allows AI shopping agents to directly access Riskified’s Identity Graph. Through this capability, merchants can retrieve risk indicators and verify customer identities programmatically. The integration can be implemented through multiple channels, including the company’s MCP integration AI Agent Approve on AWS Marketplace, Google’s Agent-to-Agent (A2A) protocol, or traditional RESTful APIs. In this setup, Riskified functions as a trusted intelligence provider, delivering real-time risk insights during critical moments of the shopping experience.
For instance, AI shopping assistants can access these signals during live customer conversations to determine whether an instant refund or product exchange request is legitimate.
In addition, Riskified has enhanced its AI Agent Policy Builder within the Riskified Decision Studio platform. This feature allows merchants to apply customized business rules that control order volume and risk levels associated with transactions initiated by AI shopping assistants. As a result, businesses can confidently deploy conversational AI agents without exposing themselves to risks such as refund fraud, reseller arbitrage, or promotion abuse.
By focusing on identity verification and behavioral anomaly detection, Riskified takes a practical approach to protecting ecommerce ecosystems. As fraud networks continue experimenting with automated systems and early agentic protocols, the company aims to ensure that AI-powered shopping remains a revenue-generating innovation rather than a security liability.
To join our expert panel discussions, reach out to info@intentamplify.com
Recommended News
Vodafone Germany extends its partnership with Amdocs to advance its multi-year transformation, adopting GenAI, cloud migration, and autonomous operations.
Arkose Labs launches an enhanced AI-powered Device ID within the Titan platform, helping enterprises detect fraud, recognize devices accurately, and prevent cyber threats.
