Diagnosed Oct 24, 2026, 9:30 AM
71/ 100
The basics are there, but some checks still keep AI from reading your information accurately or using it in answers.
Your site as AI sees it
Based on what your site declares (robots.txt and the like). It does not count actual visits.
By technical theme — points to gain
Your total can go up by 29 more points. Themes are listed by how much fixing them would add.
Content+16 pts
- Concrete figures+14 pts→ how to fix
- Internal links+1 pts→ how to fix
- Business stated up front+1 pts→ how to fix
Schema+9 pts
- Entity connections+3 pts→ how to fix
- Organization identifiable+3 pts→ how to fix
- Schema.org types+2 pts→ how to fix
- Open Graph metadata+1 pts→ how to fix
AI files+2 pts
- Markdown content negotiation+2 pts→ how to fix
Readability+2 pts
- Heading hierarchy+1 pts→ how to fix
- Landmark elements+1 pts→ how to fix
Crawlers
Every diagnosis — score over time, by stage
- Found
- Read
- Understood
- Used
Can AI find you?
Found
19/ 19100%
Same as last time (9/28)
- GPTBot—
14 passed · 4 not judged
- Applebot-Extended—
- Claude-SearchBot
- ClaudeBot—
- Google-Extended—
- OAI-SearchBot
- PerplexityBot
- Canonical URL
- Content-Signal
- /llms-full.txt present
- /llms.txt present
- llms.txt spec compliance
- Other AI crawlers—
- RSS / Atom feed
- sitemap.xml
- ApplebotN/A
- bingbotN/A
- GooglebotN/A
- MCP server declaration (/.well-known/)N/A
Can AI read your pages?
Readable
25/ 2986%
Same as last time (9/28)
- Markdown content negotiation−2 pts
9 passed · 7 not judged
- Claude-User
- Served to the crawler—
- Heading hierarchy−1 pts
- Image alt coverage
- JavaScript dependence
- Landmark elements−1 pts
- Language attribute
- Server response time
- robots.txt
- Public API referenceN/A
- ChatGPT-UserN/A
- Google-AgentN/A
- Perplexity-UserN/A
- WebMCP Declarative APIN/A
- A2A Agent CardN/A
- UCP profileN/A
Does AI understand your business?
Partly understood
27/ 3871%
+6 since last time (9/28)
- Entity connections−3 pts
- Organization identifiable−3 pts
- Business stated up front−1 pts
- Internal links−1 pts
- Open Graph metadata−1 pts
- Schema.org types−2 pts
7 passed
- Information structure
- JSON-LD present
- Structured data vs. page text—
- Meta description
- Title
- Required properties
- Property value format
Will AI use you in its answers?
Thin on grounds
0/ 140%
Same as last time (9/28)
- Concrete figures−14 pts
0 passed · 5 not judged
- Shaped like answersN/A
- Author and datesN/A
- Sources and quotationsN/A
- FAQ / Q&A contentN/A
- Recently updatedN/A
All four AI companies can find and read this site — nothing is blocked in robots.txt. Site-wide AI readiness →
Fix this first: Organization identifiable — worth 3 point(s) back.
Make the organization identifiable
Extend the Organization entity with an address, a contact and a sameAs link to an external profile. Fill the bracketed values.
JSON-LD (in a script tag)
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "サンプル株式会社",
"url": "https://www.example.co.jp/",
"address": {
"@type": "PostalAddress",
"addressRegion": "（都道府県）",
"addressLocality": "（市区町村）",
"streetAddress": "（番地）"
},
"telephone": "（電話番号）",
"sameAs": ["（会社を指す外部ページの URL。法人番号公表サイト・業界団体・SNS など）"]
}Why, and how to check
- Why
- Without an address AI cannot answer "can I use you from here". Without sameAs it cannot tell you apart from a company with a similar name (Schema.org Organization).
- How to check
https://search.google.com/test/rich-resultsRun this once the change is live. Then run the page through Tooru again to see the stage move.
- What changes
- AI understands your business better. Worth 3 point(s).
Other fixes (9)
Can AI read your pages? (3)
Give the page one h1 and a clean heading order
Exactly one h1 that says what the page is, then h2 for each topic and h3 beneath it. Do not skip levels (h2 → h4).
HTML (in the head)
<h1>サンプル株式会社 | 工場の省エネ設備の設計・施工</h1>
<h2>（主要な話題 1）</h2>
<h3>（その小見出し）</h3>
<h2>（主要な話題 2）</h2>Why, and how to check
- Why
- Agents read the outline before the body. A missing h1 or skipped levels leaves them without a map of the page.
- What changes
- AI can read more of the page. Worth 1 point(s).
Mark up header, nav, main and footer
Wrap the page in header, nav, main and footer elements, and the body itself in article or section. Replace generic div wrappers where they play those roles.
HTML (in the head)
<header>（サイト名・ナビ）</header>
<nav>（主要なリンク）</nav>
<main>
<article>（このページの本文）</article>
</main>
<footer>（会社情報・連絡先）</footer>Why, and how to check
- Why
- Landmarks tell a machine which part is the content and which is navigation. Without them the whole page is one undifferentiated block.
- What changes
- AI can read more of the page. Worth 1 point(s).
Return Markdown for Accept: text/markdown
Serve a Markdown version of the page when the request asks for it.
Why, and how to check
- Why
- It lets an agent take the structured text without parsing HTML. Support is still uneven, so treat this as a small bonus.
- What changes
- AI can read more of the page. Worth 2 point(s).
Does AI understand your business? (5)
Describe the page with more than one schema.org type
Add JSON-LD with Organization, WebSite and WebPage at least, and the type that matches the page (Article, Service, Product, FAQPage, LocalBusiness…). Put them in one @graph.
JSON-LD (in a script tag)
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@graph": [
{ "@type": "Organization", "@id": "https://www.example.co.jp/#org", "name": "サンプル株式会社", "url": "https://www.example.co.jp/" },
{ "@type": "WebSite", "@id": "https://www.example.co.jp/#site", "url": "https://www.example.co.jp/", "name": "サンプル株式会社", "publisher": { "@id": "https://www.example.co.jp/#org" } },
{ "@type": "WebPage", "@id": "https://www.example.co.jp/#page", "url": "https://www.example.co.jp/", "name": "サンプル株式会社 | 工場の省エネ設備の設計・施工", "isPartOf": { "@id": "https://www.example.co.jp/#site" } }
]
}
</script>Why, and how to check
- Why
- Each type answers a different question — who you are, what this page is, what it offers. One type answers one.
- What changes
- AI understands your business better. Worth 2 point(s).
Say what you do up front
Rewrite the title, description and h1 so that what you offer, who it is for and where you operate are all present. Fill the bracketed parts.
HTML (in the head)
<title>サンプル株式会社 | （何を）を（誰に）に提供（どこで）</title>
<meta name="description" content="（どこで）の（誰に）向けに、（何を）を提供しています。（強み・実績を 1 文）">
<h1>（何を）</h1>Why, and how to check
- Why
- AI reads the opening of the page first. If the business is not clear there, structured data cannot make up for it.
- What changes
- AI understands your business better. Worth 1 point(s).
Connect the entities with @id
Give each entity an @id and reference it from the others (publisher, isPartOf, author) instead of repeating the data. Put them together in @graph.
JSON-LD (in a script tag)
"@graph": [
{ "@type": "Organization", "@id": "https://www.example.co.jp/#org", "name": "サンプル株式会社" },
{ "@type": "WebPage", "@id": "https://www.example.co.jp/#page", "publisher": { "@id": "https://www.example.co.jp/#org" } }
]Why, and how to check
- Why
- Linked entities let a machine see that the page, the site and the company are one thing, not three unrelated records.
- What changes
- AI understands your business better. Worth 3 point(s).
Link to your other pages
Add at least five links to related pages on the same site — services, company profile, contact, related articles — as text links in the body or a related-pages block.
HTML (in the head)
<section>
<h2>関連するページ</h2>
<ul>
<li><a href="https://www.example.co.jp/（サービスのページ）">（サービス名）</a></li>
<li><a href="https://www.example.co.jp/（会社概要）">会社概要</a></li>
<li><a href="https://www.example.co.jp/（お問い合わせ）">お問い合わせ</a></li>
</ul>
</section>Why, and how to check
- Why
- Internal links are how a crawler finds the rest of the site from this page and understands what it belongs to.
- What changes
- AI understands your business better. Worth 1 point(s).
Add the four Open Graph tags
og:title, og:description, og:image (an absolute URL) and og:url in the head.
HTML (in the head)
<meta property="og:title" content="サンプル株式会社 | 工場の省エネ設備の設計・施工">
<meta property="og:description" content="サンプル株式会社は、工場や倉庫の空調・照明・コンプレッサーの省エネ設備を設計から施工、保守まで一貫して手がけています。">
<meta property="og:image" content="https://www.example.co.jp/（画像）.jpg">
<meta property="og:url" content="https://www.example.co.jp/">Why, and how to check
- Why
- They are the card an agent shows when it cites the page. Missing tags mean a bare link.
- How to check
curl -s https://www.example.co.jp/ | grep -i "og:"Run this once the change is live. Then run the page through Tooru again to see the stage move.
- What changes
- AI understands your business better. Worth 1 point(s).
Will AI use you in its answers? (1)
Put concrete figures on the page
Add prices, lead times, capacity, headcount or track record — with units. Aim for ten or more figures on the page.
Why, and how to check
- Why
- Research on generative engines (Aggarwal et al., KDD 2024) found that adding verifiable statistics raised how often a source was cited by 30 to 40 per cent. Claims without numbers are rarely quoted.
- What changes
- AI is more likely to use you in an answer. Worth 3 point(s).