Google’s John Mueller said on the latest episode of Search Off the Record that pages marked “crawled - currently not indexed” in the Search Console index report can point to a quality problem with the site, not a technical defect. Co-host Martin Splitt asked Mueller directly whether that status is often, or only occasionally, tied to quality. Mueller’s answer was blunt: “Sometimes.”
Mueller described the underlying mechanism. When Google’s ranking systems lose confidence in a site, they narrow how many of its pages get indexed and slow how often they get crawled. The practical result is a growing share of URLs stuck at “crawled not indexed” or “discovered not indexed” in Search Console, even when nothing about the page’s markup, sitemap, or server response has changed.
That framing matters because it inverts how most SEOs first read the report. A spike in unindexed URLs usually sends a technical team hunting for a broken canonical tag, a robots directive, or a crawl-budget ceiling. Mueller’s guidance suggests that hunt is often the wrong one: once a broad pattern shows up across many pages with no technical explanation, the more useful question is whether the content itself earns a place in the index.
Self-assessment is the hard part. Mueller acknowledged that judging your own site’s quality objectively is difficult, particularly for teams close to the content. He pointed to a specific failure mode: a site built largely on AI-generated pages that performed well for a stretch, until readers, and Google’s systems, start reacting the same way: “anyone could have written this, there’s nothing unique or valuable here.” He was explicit that this is not a blanket verdict against AI-assisted writing; the issue is originality and substance, not the tool used to produce the draft.
Search Engine Roundtable, run by Barry Schwartz, reported on the exchange after Google published the podcast episode. The distinction Mueller draws separates two categories of publisher risk that get conflated constantly: a crawl or rendering fault, which is fixable in days, and an accumulated trust deficit, which is not.
Sites publishing at high volume, especially those leaning on AI-assisted production, should treat this as a prompt for a specific audit over the next quarter. Pull the indexing report segmented by publish date and content type, then check whether “crawled not indexed” concentrates in a particular template, category, or production batch rather than spreading evenly. If a cluster of AI-assisted pages accounts for a disproportionate share of the unindexed set, the fix is not a new sitemap ping. It is either a rewrite that adds reporting, data, or a point of view the source material lacks, or removal of the pages entirely, since a large low-value cluster can drag down how Google evaluates the rest of the domain.
Teams that skip this step and keep resubmitting sitemaps against a quality problem will burn a quarter chasing a symptom while the underlying trust gap widens.
Search Engine Roundtable (Barry Schwartz), July 16, 2026, reporting on Google’s Search Off the Record podcast with John Mueller and Martin Splitt.
