Where it comes from
Three sources. The answers where you were absent, read for which pages the engines cited instead, which become the sites to be on. The reasons the engines gave for the rivals they recommended, which become the pages to write and the claims to make. And the site check, whose failing checks come with the fix for your platform.
How it is ordered
By what moves the answer soonest. A crawl-access failure comes first, because an engine that cannot read the site cannot recommend it. Then the pages the engines already read for your category, where a listing is a sentence away. Then the pages to write, each named for the question it answers. Housekeeping last.
What each item carries
- The question it answers and the engines that asked it.
- The sentence an engine used for the rival it named instead.
- The page it cited, so you can see what it read.
- For a check, the code to paste, written for your platform, and whether the fixed page has been read since, where the crawler sensor is installed.
Reading it from your own assistant
The plan is on the API and the MCP server, so a developer or an assistant such as Claude Code can read it and make the site changes directly. See the API and MCP reference.