Xnurta Launches MCP for Retail Media
New MCP connection brings Xnurta data and intelligence into the AI assistants advertisers already use, with a roadmap from analysis to governed action.
Topics
Xnurta, the agentic AI-powered advertising platform, has announced the general availability of Xnurta MCP, a Model Context Protocol (MCP) server that connects Xnurta retail media data with AI assistants and enterprise AI platforms, including ChatGPT and Claude.
The launch is the first phase of a roadmap to move AI assistants from answering retail media questions to taking governed actions. Xnurta’s first write capabilities, allowing AI assistants to make changes through Xnurta, enter closed beta alongside its read-only Data Query functionality.
Differentiations: Objective Context and Permission As Persistent Policy
As MCP becomes a standard interface for advertising platforms, connecting AI to advertising data is becoming table stakes. Most advertising APIs are organised around campaigns, ad groups, targets and keywords. AI can see those pieces, but does not inherently understand how they fit together around broader business goals.
ALSO READ: Why Retail Media is Moving Beyond Attribution to Incrementality
Xnurta adds that context. Its platform organises campaigns into managed groups, product lines and shared objectives that Xnurta MCP can understand. Instead of identifying every campaign tied to a product launch, an advertiser can ask how the launch is performing against its target.
As write capabilities roll out, the AI will be able to make changes across that group as a single operation.
“Marketers are already moving more of their work into general-purpose AI assistants. We don’t think they should have to move that work back into another interface to manage retail media,” said Kashif Zafar, CEO of Xnurta. “Today, Xnurta MCP starts with answers. Increasingly, it will allow those assistants to take action within boundaries that advertisers control.”
For advertisers already using AI for reporting and analysis, Xnurta MCP eliminates manual data exports. Customers can query advertising performance, business metrics, campaign structures, managed groups and product lines, along with a unified log of human and AI-initiated account activity.
ALSO READ: How Full-Funnel Retail Media is Starting to Take Shape in Practice
As Xnurta expands MCP from analysis to execution, it is introducing action spaces: configurable boundaries that determine what an AI assistant can change. This establishes permissions as a persistent policy rather than requiring approval of every AI-generated action.
