Ayudante adds MCP to MATSUBA for ChatGPT, Claude analysis

MATSUBA was launched in June 2026 as a way for advertisers to hold measurement data in their own environment with fewer gaps. Ayudante says ad platforms' management screens, GA4, and actual sign-up or purchase counts often fail to match, partly because of cookie regulations such as ITP. When conversions fall, telling a market change apart from a measurement gap has required specialist knowledge and SQL, leaving the actual cause unconfirmed, according to the company.

The new MCP feature adds a link between that stored data and AI tools. MCP is described as a common mechanism for AI tools to connect to external data, and through it ChatGPT or Claude can analyze MATSUBA measurement data accumulated in the advertiser's BigQuery. Ayudante says commonly used aggregations are prepared in advance, so staff can simply ask questions in everyday language.

Ayudante also points to broader combinations once MATSUBA data is joined with other company data in BigQuery, such as ad cost data, GA4 data, CDN access logs, and CRM or CDP sales and customer data. Because MATSUBA also sends supplemented conversions back to ad platforms, the company argues that AI bidding can learn from data closer to reality and the results can be checked again through MCP.