Fix guides
Every SeenLens check has a fix guide. Each guide says what the check tests, how to confirm the problem, how to fix it on common platforms and how to verify the fix. The paid report links each issue it finds to the matching guide.
By SeenLens · Last updated
How the guides are organised
Guides follow the five dimensions of the SeenLens score: AI crawler access, content structure, structured data, agent readiness and technical delivery. Where two checks have the same fix, they share one guide, and each check has its own section on that page.
Fix the problems that block access first. A page that AI crawlers cannot load gains nothing from better headings or structured data. The order below starts with those blocking problems.
Pages in this section
- Crawlers cannot load the page
A plain HTTP request for your page fails or returns an error, so AI crawlers have nothing to read.
- AI crawlers blocked by a firewall
A firewall or bot wall challenges or blocks AI crawlers, so they never see your page.
- Content needs JavaScript
The main content appears only after JavaScript runs, so crawlers that skip JavaScript see an empty page.
- robots.txt missing
The /robots.txt file returns 404 or a server error, so crawlers get no policy or treat the site as closed.
- robots.txt blocks AI crawlers
Your robots.txt disallows AI search crawlers or user triggered agents, which can remove you from AI answers.
- No Content-Signal line
robots.txt has no Content-Signal line, or the line is malformed or sits outside a group.
- noindex or nosnippet on the page
A noindex, nosnippet or max-snippet:0 directive keeps the page out of search and AI answers.
- Page not served over HTTPS
The page's final URL uses http://, so browsers warn visitors and engines prefer HTTPS sources.
- Soft 404 on unknown URLs
Unknown URLs on your site return 200 instead of 404, which creates soft 404s and fake discovery files.
- Canonical URL missing
The page has no canonical link, or its canonical points to another host.
- Core Web Vitals and speed
Core Web Vitals or the Lighthouse performance score are poor, so the page loads and responds slowly.
- Page title missing
The page has no title element, an empty one, or one outside 10 to 70 characters.
- Meta description missing
The page has no meta description, an empty one, more than one, or one outside 50 to 160 characters.
- H1 heading missing
The page has no H1, an empty H1, or more than one H1.
- Skipped heading levels
H3 headings with no H2 break the page outline. How to check the heading hierarchy and fix it.
- Thin content
A page with very little text gives engines little to read or cite. How the word count check works and how to add real depth.
- Opening states the answer
AI answers and snippets reuse the opening lines. How the opening check works and how to write an answer-first lead.
- Publication date
Articles need a visible date and matching datePublished and dateModified. How the SeenLens date check works and how to add dates.
- Add JSON-LD
A page with no JSON-LD states no facts for machines. How to add a JSON-LD block on Next.js, WordPress, Shopify and static sites.
- JSON-LD syntax errors
A JSON-LD block that does not parse is ignored. Common causes, a script to find the bad block and how to fix it.
- Organization schema
Organization and WebSite markup states who runs a site. A real example, Google's guidance and code for common platforms.
- Open Graph tags
Link previews and some AI tools read og:title and og:description. How to add the tags and test the preview.
- Add llms.txt
How to create and serve /llms.txt on Next.js, WordPress, Shopify and static sites, what llms-full.txt is, and how little the file is used.
- Markdown for agents
Offer clean Markdown to agents through content negotiation or .md URLs. Real response headers and setups for Next.js, Nginx and Cloudflare.
- ai-catalog.json
ai-catalog.json lists agent resources such as MCP servers. What the check tests, the ARD spec status and how to fix errors or return a real 404.
- Well-known agent files
Five optional /.well-known files for agents and APIs. Publish one only if you run the service; otherwise return a real 404.
- WebMCP tools
WebMCP lets browser agents call your site's functions. Entry point, registerTool and form annotations, the draft's status and when to skip it.
- Lighthouse agentic audit
Lighthouse scores agent readiness as a fraction. How to run the category, what each audit tests and which SeenLens guide fixes it.
- Bing AI grounding blocked
ai-grounding:disallow, noarchive or nocache can keep a page out of Bing and Copilot AI answers. How to find and fix it.
- URL path is not descriptive
A URL that ends in an ID, hash, date or query string. What a descriptive slug is, how to change it with a 301 redirect and when to skip it.
- XML sitemap missing
No XML sitemap found, or the page is not listed in it. How to create one, declare it in robots.txt and submit it to Google and Bing.