Topsort Launches MCP Server to Power Agent-Operated Retail Media

New layer enables AI agents to access, analyze, and operate workflows across retail media systems

PALO ALTO, CA / ACCESS Newswire / March 27, 2026 / Topsort, an AI-native, auction-based retail media infrastructure company powering marketplaces and retailers worldwide, today announced the launch of Topsort MCP Server, a new infrastructure layer that connects Topsort systems to AI tools through Model Context Protocol (MCP), enabling agents to access data and execute workflows across retail media systems.

MCP (Model Context Protocol) is an open standard that allows AI systems to securely connect to real-world tools and data.

As retail media programs scale, teams are responsible for managing campaigns, monitoring performance, troubleshooting delivery, and optimizing outcomes across increasingly complex and fragmented systems. Traditional workflows often require manual campaign setup, multiple dashboards, and technical resources to extract insights and make adjustments.

Topsort MCP Server introduces a new model where AI agents can directly interact with retail media systems to analyze performance, execute workflows, and optimize campaigns in real time.

“Retail media teams are under pressure to do more with increasing complexity,” said Regina Ye, CEO of Topsort. “Topsort MCP Server reflects our vision for AI-native infrastructure, where agents can interact across systems and support more automated workflows while helping marketplaces scale efficiently and maintain full control.”

From Campaign Management to Agent-Operated Retail Media

Topsort MCP Server enables AI agents to access and act on live commerce, campaign, and auction data through a standardized interface, reducing reliance on manual reporting and fragmented workflows.

Through MCP Server, agents can:

Track campaign performance in real time: Analyze spend, ROAS, CTR, CVR, CPC, impressions, and clicks

Detect pacing and delivery issues: Identify campaigns that are overpacing, underpacing, or underdelivering

Investigate performance changes: Diagnose underdelivery and performance drops

Compare performance across dimensions: Evaluate results across vendors, categories, and time periods

Support campaign operations: Recommend and assist with actions such as adjusting budgets, pausing campaigns, or updating bids

Real-Time Analytics Through Topsort MCP Server

Topsort MCP Server enables teams to query live marketplace data using natural language through AI clients such as Claude, GPT, Cursor, and other MCP-compatible tools.

Instead of navigating dashboards, exporting reports, or relying on technical teams, users can ask questions directly and receive real-time insights.

For example, a user might ask:
“Why is this campaign underdelivering this week?”

The system analyzes performance and returns clear, real-time insights, helping teams quickly understand what changed and take action.

Built for the AI-Native Future of Retail Media

Topsort MCP Server is designed to sit above existing infrastructure, allowing retailers to adopt AI-driven workflows without replacing their current stack.

The launch reinforces Topsort’s mission to build AI-native retail media infrastructure where analytics, decision-making, and execution are unified into a continuous loop that AI agents can operate directly.

Topsort MCP Server is available today and can be enabled for Topsort customers by request.

For more information or to request a demo, visit www.topsort.com.

-ENDS-

About Topsort

Topsort is an AI-native monetization infrastructure company with a mission to democratize walled garden ad technology into commerce-centric retail media infrastructure for retailers and advertisers worldwide. Today, Topsort provides monetization infrastructure to leading enterprises across more than 40 countries, including Coles, DoorDash, Woolworths, and Falabella.

Contact Information

Holly Zeng
Holly.zeng@topsort.com

SOURCE: Topsort

View the original press release on ACCESS Newswire