Microsoft Advertising MCP

Microsoft's advertising MCP server gives AI clients read access to account data so campaign questions get answered from the account rather than from a pasted screenshot. An agent can explore account structure across campaigns, ad groups, ads and keywords, and pull performance metrics including spend, clicks, impressions, click through rate, average cost per click, average cost per thousand impressions, conversions, conversion rate, revenue and return on ad spend. Results can be filtered by date range and by performance thresholds, which is what makes questions like finding keywords above a cost threshold with no conversions answerable in one step. Beyond raw metrics the server supports the analysis marketers repeat every month. An agent can surface top performing products and the search queries that triggered ads, compute month over month changes in conversions or spend, and audit how a campaign is configured against platform best practices, flagging structural problems rather than only reporting numbers. Setup involves reading the configuration guide, adding the server to the AI assistant, and starting a conversation. It works with Microsoft 365 Copilot through Copilot Studio, with the Claude and ChatGPT desktop applications, and with any MCP compatible agent. The intended users are agencies, brands and in house advertisers managing paid search across several platforms who want consolidated performance analysis in the assistant they already use, without exporting reports from another interface first.