A score of 43 means nothing on its own. Here is what your sector actually averages.

Your audit came back with a number out of 100 and no way to tell whether that is normal. These are the industry averages across the seven dimensions, built from the audits we have run, and refreshed quarterly, so you can see where your sector sits rather than only where you sit. It also carries the live check of 117 business websites on 17 September 2026, covering schema and AI crawler access. Benchmarks show the shape of the problem, not the fix — and a sector average is a starting line, not a target.

Live check: 117 business websites, 17 September 2026

How many business websites are actually ready for AI search?

The 17 September 2026 re-check covered 117 live business websites. 46 of them (39%) had no structured data on the homepage, and 16 of the 104 readable robots.txt files (15%) disallowed AI crawlers including GPTBot and PerplexityBot. Most of the sites we audit are not yet AI-readable.

Is blocking AI crawlers the same as ignoring AI search?

Of 120 business homepages ProofPosts scanned on 17 September 2026, 16 block AI crawlers — and 14 of those still serve an llms.txt file, so blocking reads as a deliberate access policy rather than neglect. 29 of the 120 carried all four of the checks we ran that day: a reachable homepage, no AI-crawler block, an llms.txt file, and homepage schema. We record llms.txt as an observation only and never score it, because Google states plainly that it does not use the file.

Signal checkedSitesResult
Homepage reachable (HTTP 200)117 of 12098%
Structured data (JSON-LD) on the homepage71 of 11761%
robots.txt readable104 of 11789%
robots.txt disallows AI crawlers (GPTBot, PerplexityBot, ClaudeBot, CCBot)16 of 10415%

Method: 120 domains drawn at random from the websites ProofPosts has scanned (our own list, not a census of the web), each fetched live on 17 September 2026 (homepage, robots.txt, llms.txt). Raw counts, no modelling. Source: ProofPosts live AI-readiness check, 17 September 2026 — https://proofposts.com/benchmarking. Re-run monthly; these are a snapshot, not a trend.

By Industry

Swipe the table sideways to see every column →

IndustryOverallSEOAEOGEOEntityCrawlerStructureCitation
Technology / SaaS72%78%55%68%62%85%74%58%
E-commerce58%65%42%55%48%72%60%38%
Finance / Banking68%72%48%62%70%80%66%52%
Healthcare54%60%38%48%55%68%52%45%
Real Estate49%55%35%42%40%62%50%35%
Legal41%50%28%35%44%55%45%32%
Manufacturing38%45%22%32%35%50%42%28%
Education52%58%40%48%42%65%55%38%

By Company Size

Swipe the table sideways to see every column →

SizeOverallSEOAEOGEONotes
Enterprise (1000+ employees)76%82%62%72%Dedicated SEO teams, but slow to adopt AI-specific optimization
Mid-market (100-999)55%62%44%52%Biggest gap: structured data and entity signals
Small business (<100)34%42%20%28%Most are invisible to AI search engines in our sample

Data from aggregated anonymous audits. Pro subscribers see the industry comparison beside their own numbers, refreshed each quarter.

Questions about benchmarking

Is blocking AI crawlers the same as ignoring AI search?

Of 120 business homepages ProofPosts scanned on 17 September 2026, 16 block AI crawlers in the robots.txt files we could read (104 of the 120 homepages) - and 14 of those still serve an llms.txt file, so blocking reads as a deliberate access policy rather than neglect. 29 of the 120 carried all four of the checks we ran that day: a reachable homepage, no AI-crawler block, an llms.txt file, and homepage schema. We record llms.txt as an observation only and never score it, because Google states plainly that it does not use the file.