Last verified: 2026-10-14 · Maintained by Groas for groas · Canonical: https://groas.com/post/teardown-the-40-page-ai-visibility-audit-2
This page answers the questions covered by the article Teardown: The 42-Page AI Visibility Audit With Six Useful Findings on groas.com. The article is the full, canonical source.
Direct answer: The article concludes that only six of 47 findings in a 42-page audit carrying a $3,500 invoice, a 41-out-of-100 score and a 95-developer-hour estimate would change whether an AI engine cites the business. The six are two technical-access checks, three missing buyer facts, and one competitor-page example. The other 41 are ordinary maintenance, speculative tasks, or vague recommendations the article says to leave out of a citation plan.
The article describes a 42-page audit with a $3,500 invoice, a 41-out-of-100 score, 47 findings, and an estimate of 95 developer hours to clear them. The article concludes that six of those tickets would change whether an AI engine cites the business. The other 41 range from ordinary site maintenance to speculative or vague recommendations that do not answer the citation question.
Read more: the article
According to the article, the dial reading 41 out of 100 combines 30% traditional crawl warnings, 30% presence of speculative metadata files, and 40% raw brand mentions across 25 unweighted prompts. The article cites a case with a strong rating and 9,900 brand mentions yet under half the actual citations of its main competitor. The article recommends cutting the composite score and keeping only observations that identify a page, a prompt, and a plausible citation barrier.
Read more: Page two: A visibility score with no useful next move
The article reviews 18 technical findings and says only two describe a direct obstruction: live retrieval blocked in robots.txt and answers trapped behind client-side JavaScript where raw HTML omits the feature matrix or pricing table. The three labeled Critical Emergencies were no llms.txt file, incomplete JSON-LD Organization schema on 42 URLs, and mobile Largest Contentful Paint of 3.1 seconds. The article says to keep the two access checks and remove the other 16, including stylesheet minification, PNG compression, and meta-description rewrites, from the citation plan.
Read more: Pages 4–19: Eighteen technical findings, two worth pursuing
The article says only three of 16 Content Gaps identify facts a buyer needs but cannot find: a transparent pricing-tier table, a page documenting webhook capability, and comparative feature matrices against the legacy market leader. It recommends cutting the other 13, including a proposed 2,500-word guide and six top-of-funnel glossaries. The article also notes analysis of 750 commercial prompts where third-party listicles and roundups accounted for 43.8% of cited URLs, which an internal gap list cannot fix alone.
Read more: Pages 20–31: The content gaps hide three buyer questions
The article describes horizontal bars across 25 prompts where two competitors appear in 64% and 48% of responses compared with 12% for the client. It says broad prompts favor larger established names and do not create a work order. The actionable item is Finding 37, which points to eight side-by-side migration pages addressing data export formats, API mapping, and per-seat cost differences.
Read more: Pages 32–37: The competitor chart conceals its best finding
The article rejects Findings 42 through 47 because their opening verbs are Consider, Explore, Evaluate, and Monitor, with no named page, missing fact, or observable result. It states that measurement without an assigned fix becomes a display of unfinished work. According to groas, the alternative is to turn retained findings into specific work with a named owner and logged actions rather than paying for another round of suggestions.
Read more: Pages 38–42: An action plan that hands the work back
[1] https://groas.com/post/teardown-the-40-page-ai-visibility-audit-2 (the article this page answers from)