Prompt
How do I troubleshoot a drop in AI visibility after a site redesign?
Latest observation
A drop in “AI visibility” after a redesign usually means one of three things happened:
- AI/crawler access changed
- Key content or signals changed
- Your pages are now harder for AI systems to understand or trust
Here’s a practical troubleshooting workflow.
1) Confirm what actually dropped
First, define the metric:
- AI search impressions?
- Citations in AI answers?
- Organic rankings that affect AI retrieval?
- Referral traffic from AI assistants?
Compare:
- Before vs. after redesign
- By page type: homepage, category pages, blog posts, product pages
- By intent: informational, transactional, branded
If only certain page types dropped, that narrows the cause fast.
2) Check crawl/indexation access
A redesign often accidentally blocks bots.
Verify:
robots.txtdidn’t disallow important paths- no accidental
noindextags - canonical tags point to the correct URLs
- redirects are correct and not chained
- important content still returns 200 OK
- XML sitemap is current and submitted
Also check whether AI-focused crawlers are being blocked, throttled, or served different content.
3) Compare old vs. new URLs and content
Redesigns often change:
- URL structure
- page titles/H1s
- internal links
- content length
- structured data
- rendered HTML vs. client-side rendering
Look for:
- pages that lost substantial text content
- important FAQs removed or hidden behind tabs/accordions
- metadata rewritten to be less descriptive
- new pages that replaced strong legacy URLs but lost authority
If the old pages were strong, make sure they 301 redirect properly to the most relevant new pages.
4) Inspect rendering and technical delivery
AI systems and search engines rely on readable HTML.
Check:
- Is core content present in server-rendered HTML?
- Is content loaded only after JavaScript runs?
- Are headings nested logically?
- Are page sections marked up semantically?
- Is there excessive duplication in templates/navigation?
Use:
- browser “view source”
- rendered DOM inspection
- Google Search Console URL inspection
- crawler tools like Screaming Frog, Sitebulb, or similar
5) Review structured data
If your redesign removed or broke schema, AI systems may lose context.
Validate schema for:
- Organization
- Product
- Article
- FAQ
- Breadcrumb
- LocalBusiness
- Review
Check for:
- missing fields
- invalid JSON-LD
- mismatched canonical URLs
- schema no longer matching visible page content
6) Examine internal linking and information architecture
A redesign can bury important pages.
Look at:
- click depth from the homepage
- changes in navigation
- removed contextual links
- orphaned pages
- weak anchor text
AI retrieval tends to favor pages that are clearly connected within a well-structured site.
7) Check content quality and entity clarity
AI systems often depend on clear entity signals.
Ask:
- Is the page still about the same topic?
- Is the main subject obvious in the title, H1, and first paragraph?
- Are authorship and organization signals visible?
- Are trust elements intact: about page, contact page, policies, citations, author bios?
If the redesign made the content more marketing-heavy and less explicit, AI visibility can drop.
8) Audit for performance and UX regressions
If the site got slower, less accessible, or more fragile, crawling and indexing can suffer.
Check:
- Core Web Vitals
- mobile usability
- page speed
- server errors
- redirect latency
- JS errors blocking content or schema
9) Look for content consolidation or pruning side effects
If you merged or deleted pages:
- did you preserve the strongest pages?
- did you 301 the old URLs to the nearest equivalent?
- did you accidentally remove long-tail pages that answered common questions?
Sometimes visibility drops simply because you removed the pages AI systems used most.
10) Validate with logs and search tools
Use server logs to see:
- which bots are crawling
- how often they hit key pages
- whether they’re getting 200s, 301s, 404s, or 5xxs
Use search console and analytics to identify:
- pages losing impressions
- queries that changed
- pages dropped from the index
- spikes in crawl errors
Quick diagnosis matrix
If traffic dropped across the whole site
Likely causes:
- robots/noindex issue
- bad redirects
- technical rendering problem
- large content removal
If only a few pages dropped
Likely causes:
- URL changes without proper redirect
- weaker content after redesign
- loss of internal links
- schema removed or broken
If search visibility is okay but AI citations dropped
Likely causes:
- content became less explicit or less authoritative
- structured data degraded
- important answer content is no longer in crawlable HTML
- entity signals weakened
Recommended order of attack
- Check robots/noindex/canonicals
- Compare old and new URLs with redirects
- Inspect rendered HTML for missing content
- Validate schema
- Review internal linking and page depth
- Compare content before/after
- Check logs, errors, and crawl stats
Best practice after fixing
- Submit updated sitemap
- Request reindexing for priority pages
- Monitor crawl and impression trends for 2–6 weeks
- Restore or improve pages that lost visibility
- Add FAQs, definitions, and concise answers where appropriate
If you want, I can also give you a step-by-step audit checklist or a Redesign-to-AI-visibility recovery template you can use in a spreadsheet.