I call BS on that. First - there's no documentation for it in Google or anywhere for that matter, second - crawlers (even with fancy AI) just don't work that way. Sitemap is essential urls that you want search engine to crawl if there are no direct connections between them (i.e. elements of the list without direct access to the list itself). Crawler job is only to get urls and follow them, after that other parts of search engine are parsing and get context for that page. That includes metadata and content, so your best chances is to put all important urls in standard sitemap, create a proper robots.txt and fill metadata (and obviously, content of each page). Everything else is AI-hype-driven speculation.