.jpg)
Xnurta, an agentic AI-powered retail media management platform, has launched The Agentic Retail Media Council, a new industry initiative to establish standardized evaluation frameworks for AI agents used in retail media and advertising. The council debuts alongside the Xnurta EVAL Framework, the first customer-facing evaluation system for retail media agents.
Xnurta launches Agentic Retail Media Council to define shared standards for evaluating advertising AI agents.
EVAL Framework measures AI agents across five dimensions: Instruction Following, Analysis Coverage, Data Accuracy, Reasoning Quality, and Recommendation Quality.
Framework evaluates agents against nearly 100 real retail media questions.
Council invites brands, agencies, and platforms to co-develop open benchmarks and audit standards.
Future EVAL iterations will be open-sourced for industry-wide use.
Xnurta also announces enhanced agent capabilities including AI optimization review and advanced diagnostics.
The Agentic Retail Media Council launches to address a critical gap: no shared definition exists for what constitutes “good AI” in retail media. The Xnurta EVAL Framework uses an open evaluation methodology to measure quality, explainability, transparency, and operational effectiveness of agentic AI systems.
The framework tests AI agents against nearly 100 real retail media questions across five dimensions: Instruction Following, Analysis Coverage, Data Accuracy, Reasoning Quality, and Recommendation Quality.
Kashif Zafar, CEO of Xnurta, said: “AI agents are already influencing optimization, recommendations, and campaign strategy across the retail media advertising ecosystem. But the industry still lacks a shared definition of what ‘good AI’ actually means. We want to take the black box and make it a glass box. The Agentic Retail Media Council exists to help create that shared definition, one built by advertisers, agencies, operators, and the broader ecosystem together.”
The council plans to collaboratively develop shared benchmark methodologies, open evaluation frameworks, AI audit and evidence standards, industry scorecards for agentic advertising systems, and an annual State of Agentic Advertising benchmark report. Future iterations of the EVAL framework will be developed collaboratively and ultimately open-sourced for industry-wide use.
Matt Yu, Product Manager at Xnurta, said: “We believe EVAL frameworks will become foundational infrastructure for trust in AI. The goal is not for one company to define the future of advertising AI. The goal is to create standards the entire industry can use to evaluate these systems consistently and transparently.”
Alongside the council launch, Xnurta announced a major update to its retail media agent capabilities, including enhanced analytical workflows, AI optimization review systems, recommendation tooling, and advanced retail media diagnostics.
About Xnurta
Xnurta is the leading agentic AI-powered retail media management platform for brands, sellers and agencies. Its AI Copilot powers over 2M bidding decisions per day based on conversion probability and retail signals on Amazon, Walmart, and Criteo, while maintaining full transparency and control over every AI-driven action. Xnurta’s AMC Hub gives marketers intuitive drag-and-drop models for path-to-conversion, customer lifetime value and cross-product association, while its Reporting Agent turns complex data into easy, shareable reports in minutes. Xnurta has won five separate Amazon Ad Partner Awards, and thousands of brands and agencies trust Xnurta to power their retail media advertising including Lenovo, Govee, Orolay, Greenworks, Jackery, Anker, and more.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.