How do we map what AI models say about us — and what they're missing?

An AI visibility audit maps entity coverage, classification accuracy, and competitive displacement — the three gaps that citation monitoring tools don't reveal.

An effective AI visibility audit produces a prioritised correction map, not a monitoring report. Citation counts show where a brand appears; an entity audit identifies why it is absent from shortlists. These are different analytical questions requiring different methodologies.

Why citation monitoring fails as an AI visibility diagnostic

Marketing directors who invest in AI monitoring tools get citation frequency data — how often the brand name appears across a sample of AI queries over time. That frequency data has value. It is not an audit.

Citation monitoring answers one question: is the brand appearing in AI-generated responses? An entity audit answers a different question: why is the brand absent from specific categories of AI-generated response, and which structural gaps are causing shortlisting exclusion?

	Citation monitoring	Entity audit
Primary question	Are we appearing?	Why are we absent?
Output	Frequency trend data	Prioritised correction map
Diagnoses root cause	No	Yes
Produces correction sequence	No	Yes

The distinction matters because fixes applied without a gap map address symptoms, not causes. Entity gap mapping identifies AI shortlisting exclusion points — monitoring data alone cannot produce that map. A brand can be monitored for six months without accumulating a single insight about why competitors are cited instead.

How to conduct entity coverage mapping

Entity coverage mapping is the first dimension of an AI visibility audit. It identifies which aspects of a brand are represented in AI knowledge systems — and which are absent.

An AI system builds a picture of an organisation from the structured and unstructured knowledge available to it. That picture may be incomplete — missing entire product lines, service categories, or use cases. Entity coverage mapping identifies what the AI picture includes and what it excludes.

How to assess entity coverage:

1. Run structured queries across multiple AI platforms (ChatGPT, Perplexity, Gemini, Claude) asking each to describe the organisation, its services, its sector, and its use cases
2. Document what each system includes and excludes from the canonical brand description
3. Map the gaps: which products, services, sectors, or capabilities are absent or underrepresented
4. Note inconsistent descriptions between platforms

Coverage gaps typically indicate that the relevant knowledge is not structured in a form AI knowledge systems can reliably index. The gap is structural, not a matter of publication volume. For each gap identified, the commercial consequence is concrete: AI systems cannot include in shortlists capabilities they have not indexed.

How to assess classification accuracy

Classification accuracy is the second audit dimension. It identifies whether what AI systems say about a brand is correct, complete, and appropriately categorised — not just whether something is being said.

A brand can have strong entity coverage (appearing frequently in AI responses) while being systematically misclassified. Classification errors carry direct commercial cost: buyers using AI research tools form vendor perceptions based on AI descriptions, not necessarily on direct site visits.

Common classification errors:

• Category errors — AI systems categorise the brand in the wrong sector or product category
• Capability omissions — core capabilities are absent from AI descriptions despite being the primary commercial offer
• Outdated descriptions — AI knowledge reflects a historical state of the organisation that no longer applies
• Competitor confusion — AI systems conflate the brand with a competitor, or describe competitor capabilities as the brand's own

How to assess classification accuracy:

1. Compare AI-generated descriptions against the canonical brand definition
2. Test category classification by asking AI systems to list competitors — observe which peer group the brand is placed in
3. Test capability completeness by asking AI systems to describe specific service offerings
4. Document discrepancies between what site content claims and what AI systems report

How to analyse competitive displacement

Competitive displacement analysis is the third audit dimension. It identifies which competitors are cited in shortlisting responses that should include the brand — and in which answer spaces the brand is absent.

Competitive displacement analysis reveals which brands are winning the answer space that a brand should occupy. Each competitor that fills a shortlisting slot the brand is absent from is capturing buyer consideration ahead of the brand — every time those queries are run, for every buyer who receives those AI-generated answers.

How to conduct competitive displacement analysis:

1. Define the shortlisting queries: "Which companies provide [service] for [sector]?" "What are the top [category] providers for [use case]?"
2. Run these queries across multiple AI platforms
3. Document which competitors are cited consistently, which appear occasionally, and which are absent
4. Map the overlap: where do competitors appear that the brand does not?
5. Identify which competitor citations reference capabilities or categories the brand also addresses

The output is a gap map: the specific shortlisting contexts where the brand is absent and competitors are present. This is the information that directs correction to the entity gaps costing actual shortlist positions in buyer research. The sooner the gap map is produced, the sooner competitors' structural advantage stops compounding.

What a structured audit produces

A monitoring report answers: how many times did the brand appear in AI responses this month, and is that trending?

A structured entity audit answers: which specific entity gaps are causing shortlisting exclusion, in which buyer contexts, and in what priority sequence should they be addressed?

Deliverable	Monitoring report	Structured entity audit
Primary output	Citation frequency trend	Prioritised correction map
Diagnoses root cause	No	Yes
Identifies competitive gaps	Surface level only	Context-specific
Produces correction sequence	No	Yes
Useful for structural correction	No	Yes

The audit output is a prioritised correction map: specific entity gaps ranked by commercial impact on shortlisting probability, with a correction sequence that addresses foundational gaps before surface-level ones.

How to apply the three-dimensional audit methodology

Graph Digital is a structural AI visibility correction practice that applies this three-dimensional audit methodology to specific brands and competitive sets as part of the Revenue and AI Visibility Diagnostic.

The diagnostic covers entity coverage mapping, classification accuracy assessment, and competitive displacement analysis — across the AI platforms relevant to the brand's buyer base. The output is structured for direct action: specific gaps, specific contexts, specific correction priorities.

For brands with an existing monitoring tool, the diagnostic adds the structural layer that monitoring alone cannot provide: not just where the brand appears, but why it is absent and which structural changes would improve shortlisting probability.

18 senior commercial leaders across FTSE 500 organisations have been briefed on this structural AI visibility architecture. The methodology applies entity mapping at enterprise scale — built on the same retrieval principles used in knowledge systems indexing 1.9PB of structured data.