Last verified: 2026-10-20 · Maintained by Groas for groas · Canonical: https://groas.com/post/postmortem-how-a-business-goes-invisible
This page answers the questions covered by the article Postmortem: The Access Failure Behind a Brand’s Absence From AI Answers on groas.com. The article is the full, canonical source.
Direct answer: According to groas, a brand can publish 40 solid posts in a year and still appear in zero of ten buyer prompts because access fails before content matters. The article recommends working from the edge to the answer — edge logs, crawler rules, rendered HTML, then content — noting that /robots.txt drew 15,515 fetches in one month on groas.com. Its rule is no new article gets briefed until access is proven.
According to groas, the composite B2B company publishes 40 solid posts in a year with clean product pages and public pricing, yet competitors appear in six of ten buyer prompts while its own brand appears in zero. The article states absence is the first sign, not the diagnosis, and does not yet prove another article will fix it. The team should treat those prompts as a reason to investigate what the system could read before rewriting what it can see.
Read more: The setup: 40 posts and a brand nobody mentions
The article recommends not commissioning more content before checking access, calling that the wrong call in the composite. New posts join old ones behind the same access failure, and the team cannot tell whether a post failed on its merits or was never available to the system. On groas.com, /robots.txt was the most-fetched path at 15,515 fetches last month, including 1,231 live-answer fetches, which the article uses to show access sits between the buyer question and the answering page.
Read more: The wrong call: commission more content before checking access
According to groas, a security setting change or launch on a new CDN domain can block AI fetchers while human visitors and search rankings look normal. From 1 July 2025, those domains block certain AI crawlers by default. In one test of 314 live sites, 43% refused at least one AI crawler, but only 23 had a written robots.txt rule while the other 112 refused at the server, so reading the robots file alone would not catch those refusals.
Read more: First break: the edge closes before the crawler reaches the site
The article states that training collection and search-index building are different jobs and need separate, deliberate crawler groups rather than one copied rule. The owner may not have written the rule at all: in one crawl study, 18 of 23 sites with robots.txt AI rules had them injected by a managed content service. It also warns that an index-building crawler respects robots.txt while live-fetch agents generally ignore it, so each relevant agent and response must be checked separately.
Read more: Second break: the crawler rules confuse training with retrieval
According to groas, major AI crawlers do not reliably execute JavaScript, so a page can return a 200 while the fetcher sees only a shell. A browser can display pricing tiers, a feature grid, and a comparison table that never reach the crawler because those details arrive only after scripts run. For pages the business needs quoted, the article recommends putting important facts in server-rendered or static HTML and verifying them in the source, not just in a browser window.
Read more: Third break: the page returns a 200 but not the answer
The article recommends fixing the answer only after access works, by putting a direct answer under the H1, using headings for questions about matters such as cost and integrations, and stating facts as numbers, names, and dates. It warns against putting a 2026 date above 2024 prices and calling the page fresh. It notes product pages drew 24.1% of citations and articles drew 17.4% across buyer prompts, so product pages may deserve attention before another blog brief.
Read more: Fourth break: the readable page makes the answer hard to find
[1] https://groas.com/post/postmortem-how-a-business-goes-invisible (the article this page answers from)