Riskified expands its AI Agent Intelligence platform to help retailers secure AI shopping assistants and prevent fraud in conversational commerce.
Riskified, a provider of e-commerce fraud and risk intelligence solutions, announced an expansion of its AI Agent Intelligence platform to help retailers securely deploy conversational AI shopping assistants on their digital storefronts.
As merchants increasingly integrate generative AI agents and chatbots into online shopping experiences, Riskified says its technology will provide a layer of risk intelligence designed to detect fraud and protect transactions occurring through these emerging interfaces.
The move reflects a broader shift across the retail sector. According to research from McKinsey & Company cited by the company, 82 percent of retail organizations have already launched generative AI pilots, many focused on reinventing customer service through conversational tools.
Also Read: The End of the Predictable B2B Buyer Journey
Securing the Next Generation of E-commerce
Retailers are exploring AI-powered assistants capable of guiding shoppers through product discovery, managing loyalty programs and processing returns or refunds. While these systems promise greater personalization, they also create new entry points for fraud and abuse.
Riskified’s platform analyzes purchase histories across a global network of ecommerce brands, enabling merchants to augment their own customer data with broader behavioral insights.
“Merchants launching virtual shopping assistants have the advantage of maintaining direct, personalized relationships with their customers,” said Assaf Feldman, co-founder and chief technology officer of Riskified. “Our role is to serve as the risk intelligence layer that both enhances and secures AI-driven interactions.”
New Capabilities for AI Shopping Agents
The expanded platform introduces several features designed specifically for conversational commerce.
AI Agent Identity Signals allow a merchant’s AI assistant to query Riskified’s Identity Graph in real time to retrieve risk indicators and verify customer identities during transactions. The system can be integrated through multiple frameworks, including cloud marketplaces, emerging agent-to-agent protocols and standard application programming interfaces.
These capabilities enable AI agents to evaluate customer risk during live conversations — for example, determining eligibility for instant refunds or exchanges.
The company also announced enhancements to its AI Agent Policy Builder, which allows merchants to establish business rules governing transactions generated through AI assistants. The feature is designed to help retailers manage risks such as refund abuse, promotional exploitation and reseller arbitrage.
Also Read: From AI Vision to Retail Personalization at Scale
Preparing for AI-Mediated Commerce
Riskified says the expansion addresses an emerging reality: fraud groups are already experimenting with early agentic commerce protocols and chatbot interfaces.
By focusing on identity verification and anomalies in purchasing behavior, the company aims to ensure that AI-powered shopping experiences remain a growth driver for retailers rather than a new source of vulnerability.
As conversational commerce continues to evolve, the challenge for retailers may be less about deploying AI assistants — and more about ensuring those assistants can distinguish legitimate customers from increasingly sophisticated digital fraud.
