AI Search Incident Diagnostic

Trace what changed before AI visibility declined.

Compare technical evidence from before and after a migration, redesign or plugin update—then rank plausible causes by impact, evidence strength and timing.

Before-versus-after evidenceLocal browser analysisNo false causation claims
01

Define the incident

Timing strengthens or weakens a hypothesis, but it does not prove causation.

Affected surfaces
02

Compare page and policy snapshots

Paste only evidence you have. Blank pairs are excluded from the diagnosis.

BEFORELast known healthy state
AFTERState after the change
03

Add optional bulk URL evidence

Use this for migrations or multi-page losses. Headers are recognized automatically.

AI Visibility Incident Diagnosis

Find what changed before AI visibility declined — without confusing correlation with causation.

The AI Visibility Loss & Technical Change Analyzer compares before-and-after technical evidence around a migration, redesign, plugin update or other website change, then ranks plausible causes according to timing, impact and the strength of the evidence you provide.

Quick Answer: Why can AI visibility drop after a website change?

AI visibility can decline after a technical change when important pages become harder to crawl, index, consolidate or discover. Potential causes include changed HTTP responses, broken redirects, robots.txt restrictions, accidental noindex directives, incorrect canonicals, outdated sitemaps, lost internal links, changed URLs or altered page content. However, a decline that happens after a technical change does not prove the change caused it. AI-platform updates, competitor gains, prompt variability, authority changes and other external factors can produce similar timing.

Visibility-loss investigations often begin with a dangerous assumption:

“We changed the website on Monday and AI mentions dropped on Friday, therefore the deployment caused the decline.”

Timing matters, but timing alone is weak evidence.

A stronger investigation asks whether the deployment created a technical change capable of affecting the affected URLs, whether that change appeared before the decline, and whether the pattern is consistent across the pages or platforms involved.

What the analyzer compares

  • Technical change date
  • Visibility decline date
  • AI mentions before and after
  • AI referrals before and after
  • Affected AI platforms
  • HTTP status before and after
  • Robots.txt before and after
  • XML sitemap URLs before and after
  • Page HTML before and after
  • Canonical changes
  • Internal-link changes
  • Bulk URL migration evidence
Diagnostic confidence is not causal proof. The tool ranks hypotheses from the evidence you supply. It does not observe AI-platform algorithm changes, competitor improvements, external authority changes or every prompt variation that could affect visibility.
Build the Timeline First

A technical incident needs a sequence of events, not just two screenshots.

The closer a technically relevant change is to a measurable decline, the more worthy it becomes of investigation. But proximity should increase scrutiny — not automatically establish blame.

Healthy State Last known period with normal visibility
Change Deployed Migration, redesign, plugin or technical update
Discovery Window Crawlers and platforms encounter the changed state
Decline Detected Mentions, citations, referrals or page visibility fall

Does a visibility drop immediately after a migration prove the migration caused it?

No. A close timeline increases the plausibility of a migration-related cause, but the hypothesis becomes much stronger only when you can identify a relevant technical difference — such as broken redirects, changed canonicals, blocked crawling, lost internal links or missing URLs — that affects the same pages experiencing the decline.

How to Use the Analyzer

Diagnose the incident in six evidence-led steps.

Add only evidence you actually have. Blank before-and-after fields are excluded rather than being treated as proof that nothing changed.

01

Define the incident

Enter the affected page, deployment date and the date when visibility loss became apparent.

Use the most precise dates available.

02

Add visibility evidence

Enter measurable before-and-after signals such as AI mentions or referral sessions where you have them.

Avoid inventing values merely to complete the form.

03

Select affected surfaces

Mark the AI platforms where the decline was observed so you can distinguish a broad visibility issue from one platform behaving differently.

04

Add before-and-after snapshots

Compare HTTP status, robots.txt, sitemap URLs and page HTML from the healthy state with the post-change state.

05

Add bulk URL evidence

For migrations or multi-page declines, upload URL-level data covering statuses, canonicals, internal links, sitemap inclusion and available visibility metrics.

06

Review probable causes

Start with high-impact findings supported by direct before-and-after evidence, then validate them on the live website before implementing recovery work.

Technical Changes to Investigate

Which website changes can plausibly affect search and AI discoverability?

The highest-value suspects are changes that alter whether a page can be requested, indexed, consolidated, discovered or understood.

HTTP

HTTP Status Changes

A URL that previously returned a successful response may now redirect, return an error, be rate limited or fail at the server level.

ROB

Robots.txt Changes

A new or broader Disallow rule may prevent supported crawlers from requesting important paths.

IDX

Noindex or Robots Directive Changes

Development or staging directives can accidentally remain on production pages after redesigns or migrations.

CAN

Canonical Changes

A page may begin declaring another URL as preferred, or several previously independent URLs may suddenly consolidate toward one destination.

301

Redirect Changes

Old URLs may redirect to incorrect, irrelevant or broken destinations, or pass through unnecessary redirect chains.

MAP

Sitemap Changes

Important new URLs may be missing while old, redirected or obsolete URLs remain listed as preferred discovery targets.

LINK

Internal Linking Changes

Redesigns can remove contextual links, navigation links or other internal paths that previously made important pages easier to discover.

HTML

Page Content & HTML Changes

Templates, headings, visible copy, metadata, structured information or content delivery can change substantially during redesigns and CMS updates.

Website Migration Analysis

Why migrations create so many possible visibility-loss signals at once.

A migration can change URLs, redirects, internal links, canonicals, XML sitemaps, server behavior, crawl access and page templates in one deployment. That makes disciplined comparison essential.

Migration check Healthy pattern Potential regression
Old URL Permanently redirects to the most relevant replacement. 404, redirect loop, temporary redirect or irrelevant destination.
New URL Returns the intended successful response. Error, block, unexpected redirect or unavailable content.
Canonical Supports the intended new preferred URL. Still references an old URL or points to an unrelated destination.
Robots / noindex Production pages expose the intended crawl and indexing policy. Staging restrictions accidentally remain active.
Internal links Point directly to current preferred URLs. Continue pointing through redirects or stop linking to important pages.
XML sitemap Lists the intended current URLs. Contains obsolete URLs or omits important replacements.
Temporary visibility fluctuation is not automatically a failed migration. Search systems may need time to recrawl and process large-scale URL changes. The important distinction is whether expected transition behavior is accompanied by genuine technical errors.
Correlation vs Causation

What makes a technical cause more believable?

A useful incident report should rank hypotheses according to multiple pieces of evidence rather than naming the most recent deployment as the cause.

Evidence Diagnostic value Interpretation
Change happened before the decline Supporting Necessary for many causal hypotheses, but not sufficient by itself.
Affected page shows a direct technical regression Strong Example: 200 changed to 404, or a self-canonical changed to another URL.
Same regression appears across affected URL group Strong Pattern-level evidence is stronger than one isolated URL.
Visibility loss appears only on one AI platform Competing explanation Platform-specific changes or retrieval differences deserve investigation.
No relevant technical difference is found Weakens technical hypothesis Do not force a technical explanation merely because a deployment happened nearby.
Fix is followed by recovery across affected URLs Additional evidence Recovery strengthens the hypothesis but can still coincide with external changes.
Competing Explanations

Not every AI visibility loss comes from your website.

The tool focuses on technical evidence you supply. A responsible investigation should also consider variables outside your website that the diagnostic cannot observe directly.

AI

Platform Changes

An AI provider can change retrieval, ranking, source selection, interfaces or citation behavior without any change on your website.

COMP

Competitor Gains

Competitors may publish stronger resources, earn better references or become more prominent sources for the same questions.

QRY

Prompt Variability

Generative answers can vary between sessions, users and prompt formulations even when the underlying website has not changed.

AUTH

Authority Changes

External links, mentions, source reputation and third-party corroboration can change independently of your technical deployment.

CONT

Content Competition

Newer or more directly relevant sources may become available for the questions where your site previously appeared.

DATA

Measurement Changes

Tracking configuration, referral behavior or monitoring methodology can change the numbers even when underlying visibility changes less dramatically.

Understanding the Diagnostic Report

Use each report tab to answer a different incident question.

OVR

Overview

Summarizes the reported decline and the most important technical evidence found in the supplied before-and-after snapshots.

TIME

Timeline

Places the technical change and visibility decline in sequence so timing can strengthen or weaken different hypotheses.

WHY

Probable Causes

Ranks plausible technical explanations using the evidence supplied rather than presenting every difference as equally important.

URL

URL Changes

Highlights multi-page patterns involving statuses, canonicals, sitemap inclusion, internal links and related URL evidence.

FIX

Recovery Plan

Converts high-priority technical findings into a sequence of issues to verify, fix and monitor.

CONF

Confidence

Reflects the strength and timing of supplied evidence. It should be read as diagnostic confidence rather than statistical or causal certainty.

Recovery Workflow

What should you do after identifying a plausible technical cause?

The goal is not to reverse every recent deployment. Fix the smallest verified technical regression first, then monitor whether visibility and crawl evidence improve.

01

Verify the live issue

Confirm that the suspected status, redirect, robots, canonical, sitemap or HTML problem still exists on the production site.

02

Measure the scope

Determine whether the issue affects one URL, one template, one directory or the entire site.

03

Fix the verified regression

Correct the specific technical problem rather than rolling back unrelated design, content and infrastructure changes simultaneously.

04

Restore supporting signals

Update internal links, sitemap URLs, canonicals and redirects so they consistently support the intended page state.

05

Encourage rediscovery

Use relevant search-engine inspection and sitemap tools where appropriate, then allow time for crawlers and external platforms to encounter the corrected state.

06

Monitor recovery

Compare technical state, crawler activity, search performance, AI mentions, citations and measurable AI referrals across consistent reporting periods.

Do not roll back blindly. If a redesign changed 50 things but only the canonical template is demonstrably wrong, fixing the canonical issue creates a cleaner test than reverting the entire website.
Better Evidence Collection

Why before-and-after snapshots are valuable for technical SEO and GEO.

Technical incident diagnosis becomes dramatically easier when you can compare the last known healthy configuration with the post-change configuration.

Useful before-change evidence

  • HTTP response
  • Final resolved URL
  • Canonical annotation
  • Robots meta directives
  • Relevant response headers
  • robots.txt
  • XML sitemap URLs
  • Internal-link counts
  • Rendered or source HTML
  • Search and AI visibility baseline

Useful post-change evidence

  • Same technical fields collected again
  • Deployment timestamps
  • Redirect maps
  • Crawl errors
  • Server logs
  • Search Console changes
  • AI crawler activity
  • AI citations and mentions
  • AI referral traffic
  • Recovery-test results
SEO Loss vs AI Visibility Loss

The same technical issue can appear differently across search and AI platforms.

Observed pattern What it may suggest
Google rankings, impressions and AI visibility all decline A broad technical or content issue becomes more plausible, but still requires page-level validation.
Traditional search declines while AI visibility appears stable The issue may affect conventional search more strongly, or AI platforms may still rely on previously discovered or external information.
AI visibility declines while traditional search remains stable Platform-specific changes, source-selection changes, competitor gains or AI retrieval differences deserve greater attention.
Only one AI platform declines A universal website-level technical cause becomes less convincing unless that platform has unique access requirements.
Multiple affected URLs share one technical regression The common technical change becomes a stronger candidate for investigation.
Practical Use Cases

When this analyzer is most useful.

01

Website Migration

Compare old and new URL states after domain changes, slug changes, HTTPS migrations or site-architecture restructuring.

02

Website Redesign

Investigate whether new templates changed internal links, metadata, canonical tags, page HTML or important discovery paths.

03

Plugin or CMS Update

Compare technical output before and after SEO plugins, caching systems, page builders or CMS configurations were changed.

04

AI Visibility Decline

Test whether a reported drop in mentions, citations or referral traffic lines up with a technically plausible page-level regression.

05

Multi-Page Traffic Loss

Use bulk URL evidence to find common changes across groups of pages rather than diagnosing URLs individually.

06

Client Incident Report

Replace vague statements such as “the redesign hurt GEO” with an evidence-based timeline, ranked hypotheses and specific recovery checks.

Who Should Use It

Built for teams investigating technical visibility incidents.

Technical SEO Professionals

Compare crawl, indexability, canonical, sitemap, redirect and internal-link states around a reported traffic or visibility decline.

AEO & GEO Specialists

Test whether a decline in AI visibility has a plausible technical explanation before recommending content rewrites or authority campaigns.

SEO Agencies

Create structured incident reports for migrations, redesigns and unexpected performance losses across client websites.

Developers

Identify production changes that may have altered response behavior, metadata, links, templates or URL handling.

Website Owners

Understand whether a recent site change deserves technical investigation instead of immediately blaming content quality or an algorithm update.

Analytics Teams

Connect measurable before-and-after visibility signals with a documented deployment timeline while keeping correlation and causation separate.

Incident Diagnosis Best Practices

How to avoid bad conclusions when visibility drops.

Record exact deployment dates

“Around the beginning of the month” is much weaker evidence than a documented production deployment timestamp.

Compare equivalent snapshots

Compare the same URLs and technical fields wherever possible so differences represent real changes rather than different collection methods.

Change one thing at a time where practical

Combining domain migration, CMS migration, redesign and content rewrite into one launch makes later diagnosis much harder.

Investigate affected and unaffected pages

A technically meaningful cause should often explain why one URL group declined while another group remained healthy.

Use live validation

Historical evidence identifies suspects. Confirm the current HTTP, robots, canonical, sitemap and page state before fixing anything.

Check platform breadth

A decline across several search and AI surfaces supports a different hypothesis than a decline isolated to one provider.

Document external changes

Record known search updates, AI-platform changes, competitor activity and major external events that could provide alternative explanations.

Measure recovery with the same method

Do not compare one monitoring methodology before the fix with a completely different methodology after it.

Know the Limitations

What the analyzer can — and cannot — establish.

The tool can help you

  • Build a before-and-after incident timeline
  • Compare supplied HTTP evidence
  • Compare robots.txt snapshots
  • Compare sitemap URL lists
  • Compare supplied page HTML
  • Analyze bulk URL changes
  • Rank plausible technical causes
  • Identify affected URL patterns
  • Create a recovery checklist
  • Export findings for reporting

The tool cannot prove

  • That a website change caused the decline
  • How an AI platform changed its algorithms
  • Whether competitors improved externally
  • Every change in backlinks or authority
  • Every prompt or answer variation
  • Whether an AI provider refreshed its source index
  • Future recovery timing
  • Guaranteed restoration of AI citations
  • Guaranteed search ranking recovery

Use the tool to rank hypotheses — then try to falsify them.

A strong diagnosis is not the explanation that sounds most convenient. It is the explanation that survives live technical validation, affected-versus-unaffected comparisons and competing alternative causes.

Frequently Asked Questions

AI visibility loss, migrations and technical SEO incidents explained.

Direct answers to common questions about losing AI citations or mentions after website changes, migrations, redesigns and technical updates.

Why did my AI visibility drop after a website migration?

A migration can change URLs, redirects, canonicals, internal links, sitemaps, crawl rules, page templates and server responses. Any of these can create a plausible technical visibility problem. Compare the last known healthy state with the post-migration state before assigning a cause.

Can a website redesign reduce AI visibility?

Yes, if the redesign changes technically or semantically important elements such as URLs, internal links, crawl access, canonical tags, page content, headings or content delivery. A visual redesign by itself is not evidence of a visibility problem.

Can a WordPress plugin update cause visibility loss?

It can if the update changes technical output such as robots directives, canonical tags, redirects, sitemaps, metadata, page rendering or server behavior. Compare before-and-after evidence rather than assuming the plugin update is responsible merely because the dates are close.

Can robots.txt changes reduce AI visibility?

Restricting relevant compliant search or AI crawlers can reduce one route through which pages are discovered or refreshed. The effect depends on the crawler, affected paths and the platform's broader retrieval systems.

Can an accidental noindex cause a major visibility decline?

Yes. An unintended noindex directive can prevent supporting search engines from keeping a page indexed. Development and staging restrictions should therefore be checked carefully after migrations and redesigns.

Can a wrong canonical cause visibility problems?

A canonical pointing to an unintended URL can signal that another page is the preferred representative of the content. Canonical annotations are signals rather than absolute commands, but incorrect implementations deserve immediate investigation.

Can broken redirects cause AI visibility loss?

They can contribute to discoverability and URL-transition problems when old URLs no longer lead to the correct replacement. Check redirect destinations, chains, loops, errors and whether external references still point toward older URLs.

Can removing internal links hurt important pages?

Internal links help users and crawlers discover pages and understand site relationships. A redesign that removes important contextual or navigational links can therefore change how prominently a page is connected within the website.

Can an outdated XML sitemap cause AI visibility loss?

A sitemap is only one discovery signal, so an outdated sitemap does not automatically cause visibility loss. However, after migrations or URL changes, keeping obsolete URLs while omitting current preferred URLs can weaken the clarity of your discovery signals.

How long does visibility recovery take after a migration?

There is no universal recovery period. Search systems need time to recrawl and process changed URLs, and AI platforms have their own refresh and retrieval systems. Recovery speed depends on the scale of the change, crawl activity, technical correctness and platform-specific behavior.

Should I immediately roll back a redesign after traffic or AI visibility drops?

Usually not without evidence. First identify whether a specific regression exists. Fixing one verified technical issue creates a cleaner recovery test than reversing many unrelated changes simultaneously.

Does a decline happening after a deployment prove causation?

No. Timing can strengthen a hypothesis, but causation requires stronger evidence. Look for a relevant technical change that affects the same URLs and precedes the observed decline.

What is diagnostic confidence?

Diagnostic confidence reflects how strongly the supplied evidence supports a particular hypothesis. It is not a statistical probability and does not prove that the proposed issue caused the visibility decline.

Why might only ChatGPT or Perplexity visibility drop while Google stays stable?

AI platforms use different crawlers, retrieval systems, source indexes and ranking methods. A decline isolated to one platform can therefore point toward platform-specific behavior or competition rather than a universal website technical failure.

Can competitor improvements cause my AI visibility to decline?

Yes. AI answers have limited attention and citation space. A competitor publishing stronger information, earning better external references or becoming a more suitable source can reduce your observable visibility even when your website has not deteriorated technically.

Can AI referral traffic fall without AI visibility falling?

Yes. Referral traffic depends on users clicking identifiable links and attribution reaching analytics. Citation visibility, click behavior and analytics attribution are separate layers.

What data should I save before a website migration?

Save URL inventories, redirect mappings, HTTP responses, canonical tags, robots directives, XML sitemaps, internal-link data, representative page HTML, Search Console baselines, analytics data and any AI visibility metrics you actively monitor.

How do I know whether a technical fix worked?

First verify that the technical regression is actually corrected. Then monitor crawling, search performance, AI mentions, citations and referral data using the same methodology and comparable reporting windows used before the fix.

Can this tool guarantee recovery of lost AI citations?

No. The analyzer helps identify technically plausible causes and recovery actions. AI platforms independently control discovery, retrieval and citation selection, so correcting a technical issue cannot guarantee future citations.

What if the tool finds no meaningful technical change?

That weakens the technical-change hypothesis. Expand the investigation to content changes, authority, competitors, platform updates, prompt variability, measurement changes and other external explanations instead of forcing a technical cause.

Evidence Before Blame

A visibility decline after a deployment is a clue — not a conclusion.

Use the AI Visibility Loss & Technical Change Analyzer to compare the last known healthy state with the post-change site, rank technically plausible causes, challenge competing explanations and build a recovery plan based on evidence rather than assumptions.

Scroll to Top