Sky Media LegacySites score 15 points lower on AI readiness than on SEO
Average score out of 100 across 326 sites
Your buyers now ask ChatGPT, Perplexity and Google's AI which vendor to pick before they ever open your site. So I ran the homepages of 326 fintech, crypto and AI companies through the same checks I use for my free AI score.
The good news is almost nobody is locking AI out. The less good news is plenty of sites let AI in, but give it nothing useful to read.
Key findings
- 52% of the 326 sites scored under 70 out of 100 on AI readiness.
- The 326 sites averaged 85/100 on SEO but only 70/100 on AI readiness.
- 31% of the 156 sites with an SEO score of 90+ scored under 70 on AI readiness.
- 49% of the 326 sites scanned have no llms.txt file for AI assistants.
- 41% of the 326 sites have no structured data saying what the company sells.
- Only 4 of the 326 sites block AI crawlers, so access is rarely the problem.
- 49% of 125 crypto sites scored under 70, versus 25% of 144 AI sites.
Is good SEO enough to show up in AI answers?
No. Across the 326 sites, the average SEO score was 85 and the average AI-readiness score was 70. 199 of the 326 sites (61%) scored higher on SEO than on AI readiness.
The gap holds even at the top. Of the 156 sites with an SEO score of 90 or more, 49 (31%) still scored under 70 on AI readiness. For the industry, that means years of SEO work won't carry over on their own. For your company, a clean Google audit is not proof that AI can describe you.
The surprise is at the very top. 90 sites scored a perfect 100 on SEO, and 23 of them (26%) still scored under 70 on AI readiness.
So I decided to audit for AI specifically.
Overall scores split in two: a third under 70, a third at 90+
116 sites (36%) scored under 70; 118 (36%) scored 90 or higher
Turns out "we rank fine on Google" is the 2026 version of "we have a fax number."
Are AI crawlers blocked from your site?
Almost never. Only 4 of the 326 sites block AI crawlers like GPTBot in their robots.txt file, the file that tells bots what they may read.
That changes where the industry should look. Access is solved for nearly everyone; what's missing is content AI can use once it arrives. For your company, check robots.txt once, then spend your time on the gaps below.
Does your site tell AI what you sell?
Often not. 135 of the 326 sites (41%) have no Organization or Product schema, the structured data that labels who you are and what you sell in a format machines read directly.
Without it, AI has to guess your category from marketing copy. In crowded categories like stablecoin payments or AI agents, guessing is how you end up described as "another platform." The fix is a small block of JSON-LD on your homepage naming your company, your product and what it does.
What is llms.txt, and does it matter?
llms.txt is a plain-text file at the root of your site that summarizes what you do for AI assistants. 159 of the 326 sites (49%) don't have one, the most common gap in the scan.
To be straight with you, no major AI engine has confirmed it relies on llms.txt yet. I check for it because it's an emerging convention and it forces you to write the clear summary AI needs anyway. It's an insurance policy to make sure an AI understands what your company does.
Can AI pull a clean answer from your homepage?
For a third of sites, no. 111 of the 326 sites (34%) have weak structure: few headings, no lists and no tables for AI to quote.
AI answers favor content that comes in easy pieces, like a feature list, a pricing table or a short FAQ. A homepage that's one long scroll of slogans gives AI nothing to lift. The best outcomes come from stating what you do, who it's for and how it's priced in headings and lists.
Which vertical is furthest behind?
Crypto. On AI readiness alone, the 125 crypto sites averaged 63 out of 100, against 72 for the 57 fintech sites and 76 for the 144 AI sites. 58% of crypto sites had no llms.txt and 54% had no company or product schema.
This is a huge problem in categories (especially regulated ones) where buyers ask AI whether a platform is safe and licensed. If AI can't find clear facts on your site, it answers with whatever third parties say.
Crypto sites miss the two biggest basics most often
Share of sites missing each check, by vertical
What do the top scorers do differently?
They cover the basics without exception. 118 of the 326 sites (36%) scored 90 or higher, and 52 (16%) scored a perfect 100.
| Check | Sites scoring 90+ (118) | Sites scoring under 70 (116) |
|---|---|---|
| No llms.txt | 1% | 90% |
| No Organization or Product schema | 2% | 91% |
| Weak structure (few headings, lists or tables) | 14% | 54% |
| Average SEO score | 95 | 73 |
| Average AI-readiness score | 98 | 43 |
Yes, the score counts these checks, so missing them drags a site down. But the low scorers also trail on SEO, 73 to 95, and SEO counts neither check. The top sites aren't doing anything clever. They just wrote down what they do where machines can read it.
How I ran the scan
I scanned the homepages of 326 companies in three verticals: crypto infrastructure, fintech and B2B AI tools. The dataset is as of October 2026.
Each site gets two sub-scores out of 100. The SEO score covers the title and description, headings, language tag, sitemap, crawl instructions and HTTPS. The AI-readiness score covers AI crawler access in robots.txt, llms.txt, Organization and Product schema, and how much of the page sits in headings, lists and tables. The overall score is the average of the two.
The scan has limits. It reads the homepage only, so a strong pricing or docs page elsewhere doesn't count. It reads the HTML the server sends, so content that only appears after JavaScript runs can be missed. Page speed wasn't measured. And it measures whether a site is ready for AI, not whether AI actually recommends it.
No company is named in this report. Results are reported only in aggregate.