On the October 1, 2026 "Do sitemaps still matter?" episode of Google's Search Off the Record podcast, Mueller said his own logs show AI crawlers requesting sitemap and RSS files, and advised publishers who want to be found to use a standard sitemap name or lean on RSS feeds.
2026-10-07 · by Moe Ameen
On October 1, 2026, Google published a Search Off the Record podcast episode titled "Do sitemaps still matter?" in which Search Relations team members John Mueller and Martin Splitt discussed the current role of XML sitemaps. In it, Mueller said he has seen AI crawlers request his sitemap and RSS files in his own server logs, as reported by Search Engine Journal on October 5. He did not name which crawlers, and he said he does not know whether the AI companies document the behavior or what they do with the files once they fetch them. So this is an observation from one person's logs, not a formal Google specification of how AI systems discover or use publisher content.
Mueller's practical point was about discoverability. AI training crawlers, he noted, generally give site owners no console or setup to submit a sitemap the way Google Search Console does — they typically lack "any kind of a Console or any setup where you can submit a sitemap file." For publishers who want AI systems to find their content, he suggested two low-effort moves: name your sitemap with the generic "sitemap.xml" so a crawler guessing at the standard location finds it, or lean on RSS feeds, which are easier to discover because they are usually linked from a page's HTML head. Google's own documentation already accepts supported RSS and Atom feeds as a form of sitemap.
He drew the opposite line for anyone who wants a private sitemap: give it an unusual file name, keep it out of robots.txt, and submit it directly to Google in Search Console — Bing would likely need its own submission. He also pushed back on llms.txt, the proposed file for pointing AI models at a site's content, as a sitemap substitute, calling the hope invested in it "bigger than the reality."
Separately in the same episode, Mueller attributed "couldn't fetch" errors on otherwise-valid sitemaps to host load and crawl demand, adding that crawl demand is "very often based on the perceived quality of a website." As with any podcast remark, treat this as guidance rather than committed policy, and confirm specifics against Google's own documentation before building a strategy around a single detail.
The quiet takeaway is that the discovery plumbing only matters if there is something worth crawling behind it. A sitemap and an RSS feed are just lists of your pages; an AI crawler that finds them still has to find real, regularly updated, on-brand content on the other side. For most creators the bottleneck was never the file format — it is keeping a steady stream of original pages flowing so the feed is never stale.
That supply problem is where [Kompozy](/) fits. Kompozy generates Blog Articles, Email Newsletters, and Text Posts from your source material and publishes the blog pieces straight to WordPress, GHL Blog, or a custom webhook — the destinations that emit the RSS feed and sitemap these crawlers follow — while [Autopilot](/glossary/autopilot) fans matching social posts across the eight social platforms plus blog and email behind a per-post review. The [Persona Brief](/glossary/persona-brief) keeps every piece in your voice, so what a crawler fetches is genuinely yours, not templated filler. Kompozy won't rename your sitemap.xml or edit robots.txt — that stays your SEO housekeeping — but it handles the harder half: making sure there is a consistent, canonical body of content for an AI crawler to discover in the first place.
Not as a formal policy. Google's John Mueller said on the Oct 1, 2026 Search Off the Record podcast that he has personally seen AI crawlers request his sitemap and RSS files in his own server logs. He didn't name the crawlers or say what the companies do with the files, so it's a real observation from one set of logs rather than an official specification.
Mueller's advice was to use both. Name your sitemap with the generic "sitemap.xml" so a crawler can find it at the standard location, and keep an RSS feed, which is easier to discover because it's usually linked in the page's HTML head. Google's own documentation already accepts supported RSS and Atom feeds as a form of sitemap.
No. Mueller was clear he doesn't know what AI companies do with the files they fetch. A crawler reading your sitemap or feed is only discovery — the content on those pages still has to be original and useful enough to earn a citation or an answer.
Mueller said no, describing the hope placed in llms.txt as "bigger than the reality." He suggested publishers focus on standard sitemaps and RSS feeds rather than treating llms.txt as a sitemap substitute.