The challenge with measuring AI visibility

AI responses are not deterministic. Ask ChatGPT the same question twice and you may get different answers. Ask it with slightly different phrasing and the results can vary significantly. This makes AI visibility inherently harder to measure than a Google ranking, which is more stable for a given query, device, and location.

Our approach addresses this by focusing on two things: the structural factors that reliably influence whether a business is cited, and a consistent query methodology that tests representative patterns rather than single data points.

Platforms we audit

1
ChatGPT (GPT-4o), tested with and without web browsing enabled. We test the business as a named entity ("tell me about [business name]") and in recommendation queries ("best [trade] in [area]").
2
Perplexity AI, one of the highest-citation AI platforms, with a strong focus on sourced answers. We test local recommendation queries, cost queries, and direct entity queries.
3
Google AI Overviews, tested via standard Google search for queries where AI Overviews are triggered. We check both direct entity presence and indirect citation in sector-level answers.

The seven dimensions we score

1. Entity recognition
Does AI correctly identify what this business is, what it does, and that it exists as a real, verifiable entity?
2. Location accuracy
Does AI correctly state where the business operates and its service area?
3. Service coverage
Can AI accurately describe the services the business offers when asked?
4. Credential recognition
Does AI recognise and cite the business's professional credentials and accreditations?
5. Citation frequency
How often does the business appear in recommendation queries for its sector and area?
6. Content relevance
Does the business have content that directly answers the questions customers ask AI about it?
7. Trust signal completeness
Are reviews, credentials, and entity data consistent and complete across the web?

What we do not measure

We do not claim to predict exactly which queries a business will appear in across every AI platform "” that is not currently possible with any reliability. AI models are updated frequently, and citation patterns shift with model versions, training data, and query phrasing.

What we do measure is the structural readiness of a business's online presence for AI visibility "” and where the specific gaps are. These structural factors are the most actionable things a business can change, and they have a reliable and measurable impact on AI citation likelihood over time.

Questions about our AI visibility tracking methodology

Which AI platforms does aivisible.co.uk check in an audit?
+
Our standard audit checks ChatGPT (GPT-4o), Perplexity AI, and Google AI Overviews. These three cover the majority of AI-driven discovery queries in the UK. Our methodology tests multiple query variations and records typical response patterns rather than single data points.
How do you score AI visibility?
+
Our AI visibility score is based on seven dimensions: entity recognition, location accuracy, service coverage, credential recognition, citation frequency, content relevance, and trust signal completeness. Each is scored and weighted to produce an overall score with specific recommendations.
How reliable are AI visibility scores?
+
AI responses are probabilistic and vary between sessions. Our scores represent a snapshot of how AI systems currently interpret your business based on your online presence. We focus on the structural factors that reliably influence AI visibility rather than on individual query outputs.