AI Visibility Audit
Review public website evidence before making AI search claims.
AuditHQ reviews observable website signals such as crawler rules, structured data, entity information and content structure. Use the recorded evidence to investigate a concrete issue. A readiness finding is not proof that an answer engine has read, trusted or cited the site.
Scope note. This is a website-signal readiness review. It does not measure actual AI mentions, citations or referrals, and it does not establish how every provider retrieves or selects content.
Audit suite
AI Visibility
Verified checks
Example findings
9
Every check verified against live data.
Grouped into 5 categories. Code runs every check against your live site - AI only writes the narrative around verified findings. Methodology is proprietary; the cards below describe what each category catches.
01
14 checksStructured Data
Observed Schema.org markup and its relationship to visible information. Presence alone does not prove feature eligibility, factual accuracy or AI citation.
02
12 checksAI Crawler Access
Declared crawler policies and sampled public responses. Search crawling, model training and user-initiated retrieval have different roles; a robots.txt allowance is not proof of a provider visit.
03
11 checksEntity Clarity
Names, business information and profile references that a reviewer can compare for consistency. A sameAs link does not establish ownership, account verification or universal machine trust.
04
14 checksContent Structure
Readable text, headings, tables, sources and authorship information for human review. These are observations rather than a measured formula for answer-engine selection.
05
10 checksRendering
Initial HTML and available rendering evidence. Consumers differ in JavaScript support; the report cannot establish what every AI service sees.
Specific check definitions, regex patterns, and severity rubrics are not published. The output is a verified, evidence-backed report - not a methodology document.
Three evidence boundaries to keep in view.
Find a specific access or content problem before choosing a fix.
A blocked path, missing rendered content or contradictory business fact can be investigated with recorded evidence. Keep the affected URL, consumer and observation with the change. A missing optional file alone does not establish lost traffic.
Google requires no special AI file or schema.
Google says its AI search features need no special schema or AI text file. An optional llms.txt may serve a documented client workflow, but it is not a universal prerequisite. Structured data should describe visible facts; it cannot make every AI citation accurate.
Explain authorship without inventing a ranking requirement.
Accurate bylines, relevant experience and cited sources help readers evaluate a claim. Google describes E-E-A-T as a quality-assessment concept rather than one specific ranking factor. A named author or LinkedIn link alone is not a universal Google or AI ranking requirement.
AI Visibility signals
Six website observations to review against the intended consumer.
AI search systems can combine indexed content, live retrieval, citations, and model knowledge. AuditHQ checks observable website-readiness signals such as entity clarity, content structure, and crawler access. It does not currently claim that those checks measure your actual mention rate inside every AI answer engine.
Organization & LocalBusiness schema
Describe verified business facts where the type fits the page. Compare the markup with visible content and check the intended consumer's documentation. Missing markup does not prove that an AI service misunderstands the business; Google requires no special schema for its AI search features.
Review scope
- Visible facts
- Applicable schema
AI crawler access (robots.txt)
Record the named user agent, matching path and declared rule. Search crawling, model training and user-initiated retrieval can have different controls. A declared allowance is not evidence of a visit, and a disallow rule does not establish everything an assistant can know from other sources.
Review scope
- Declared rules
- Consumer-specific roles
llms.txt declaration
An optional proposed map of site content. Confirm a documented consumer need before implementing it. Google says its AI search features need no special AI text file; presence or absence alone does not establish eligibility, ranking or citation outcomes.
Review scope
- Optional
- Documented consumer need
Entity sameAs links
Review whether linked profiles describe the same real organisation. A declared sameAs URL is not proof of profile ownership, account verification or an answer engine's trust. Correct contradictory identity information before adding more links.
Review scope
- Identity consistency
- Ownership reviewed separately
Readable content and sources
Use clear headings, complete answers and sources appropriate to the claim. Structure should help a reader understand the page. There is no universal heading, FAQ schema or passage-length formula here that establishes extraction or citation by every engine.
Review scope
- Reader usefulness
- Source accuracy
Named-author bios with credentials
State real authorship and relevant experience when appropriate. Do not invent qualifications or require a LinkedIn profile as a universal ranking rule. Authorship information helps readers assess provenance; it does not prove how an answer engine will rank or cite the page.
Review scope
- Actual authorship
- Relevant experience
Evidence boundaries
Separate a website observation from an AI search outcome.
Website observation
Record the URL, time, sampled response or rendered page and the specific rule or content observed.
A sampled website signal does not establish a provider visit, indexing, inclusion or a citation.
Documented consumer
Identify the intended search feature, integration or crawler and check its current published requirements.
Do not apply one provider's requirements to every engine or treat optional files and schema as universal prerequisites.
Observed outcome
Measure mentions, citations or referrals separately using a defined query set, engine, geography and period.
Website readiness scores alone do not establish traffic growth or the internal selection rules of an answer engine.
The full AI Visibility suite reviews public website signals. Create an account to start a Full Audit with supporting evidence and a prioritised action plan. This reviews public signals, not observed AI-answer tracking.
Start Full AuditIllustrative finding: no Organization markup observed
Evidence
Fictional example: a sampled homepage response contains no Organization JSON-LD. This describes the response reviewed, not the business identity data available to every search or AI system.
Recommended fix
Confirm the business name and website URL visible on the homepage. If appropriate for a documented consumer, add consistent Organization markup, check for conflicting CMS output and validate the deployed page.
Business impact
The repair adds an explicit machine-readable description of verified facts. Measure any subsequent citation, referral or commercial outcome separately.
Fictional example of the evidence, fix and impact structure. It is not a current report or a finding about your site.
Questions, answered.
What buyers ask before running their first ai visibility audit.
Stop reading. Start auditing.
Audit your ai visibility stack now.
Create an account to start a Full Audit. This suite includes 116 catalogue checks; applicability depends on your site and scope. Review recorded evidence and a prioritised action plan.