Fix guide
Copy-pasteable fixes for every check in the audit, heaviest lanes first.
Every check in the audit has a concrete fix. This page collects all of them with copy-pasteable examples, ordered by lane weight — start at the top and the score moves fastest.
How do I unblock AI crawlers? (lane weight 25)
Open robots.txt at your site root and look for Disallow: / under a User-agent naming an AI bot. To allow the engines the audit checks:
User-agent: GPTBot
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: Claude-Web
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Google-Extended
Allow: /
User-agent: CCBot
Allow: /
If you have no robots.txt at all, that's fine — everything is allowed by default. Only add rules when you want to restrict something. And remember robots.txt is a request, not enforcement: it keeps honest crawlers out, nothing more.
llms.txt — serve a markdown file at /llms.txt that maps your key pages:
# Your Company
> One-sentence description of what you do and for whom.
## Product
- [Pricing](https://example.com/pricing): plans and what each includes
- [Docs](https://example.com/docs): setup and API reference
## Company
- [About](https://example.com/about): team, history, contact
Keep it short and curated — it's a reading guide for machines, not a sitemap dump.
How do I make a page citable? (lane weight 25)
- Exactly one H1. State the page's answer in plain language. Multiple H1s blur what the page is about; zero leaves engines guessing.
- At least two H2 sections. Engines cite passages, and passages need boundaries. Each H2 should mark a self-contained chunk.
- 300+ words of real text. Thin pages don't get cited. This is visible text — not markup, not alt attributes.
- Use lists. Bulleted steps and comparison tables are the formats AI answers extract most often.
- Ask the question, then answer it. At least one heading phrased the way users ask ("How much does X cost?") followed by a direct 1–2 sentence answer before any elaboration.
One structural warning: the auditor (like most crawlers) reads the HTML your server returns. If your content only renders client-side, it doesn't exist for this lane. Server-render anything you want quoted.
How do I fix metadata? (lane weight 20)
<title>Plain-language product name and who it's for (15–70 chars)</title>
<meta name="description" content="A 50–160 character answer to the query this page targets. Engines lift these verbatim, so write it as the answer." />
<link rel="canonical" href="https://example.com/this-page" />
<meta property="og:title" content="Same title, social-ready" />
<meta property="og:image" content="https://example.com/og.png" />
And serve everything over HTTPS — engines deprioritize insecure sources, and some fetchers refuse them outright.
What structured data do I need? (lane weight 20)
Minimum viable JSON-LD — an entity block so engines can verify who you are:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@graph": [
{
"@type": "Organization",
"name": "Your Company",
"url": "https://example.com",
"logo": "https://example.com/logo.png",
"sameAs": [
"https://github.com/yourcompany",
"https://www.linkedin.com/company/yourcompany"
]
},
{ "@type": "WebSite", "name": "Your Company", "url": "https://example.com" }
]
}
</script>
Then add FAQ schema for your top 3–5 real questions — the highest-yield format for AI answers:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "How does pricing work?",
"acceptedAnswer": { "@type": "Answer", "text": "Direct, complete answer here." }
}]
}
</script>
Only mark up content that's visibly on the page. Schema describing things the page doesn't say is the fastest way to get structured data ignored.
How do I fix trust signals? (lane weight 10)
- Link an About page (
/about) from the audited page — entity verification is a core trust signal. - Show a contact route: a
/contactlink or a visiblemailto:address. - Add freshness markup: a
<time datetime="2026-07-18">on visible dates, ordatePublished/dateModifiedin your JSON-LD.
In what order should I fix things?
The audit already ranks your top five by lane weight, but the general order is: unblock crawlers (nothing else matters if engines can't read you) → make the page citable → metadata → schema → trust signals. Re-run the audit after each change; scans from the same browser keep a local history so you can watch the score move.