Measurement starts with a baseline. Before any work begins, a researcher records how AI systems currently describe a business: which platforms mention it, which sources they cite, what facts they get right or wrong, and how the business compares with named competitors on the same questions. That baseline is the only honest reference point for later change. An audit then examines the underlying public record — the entity footprint described in Entities, Knowledge Graphs and Semantic SEO — checking for inconsistencies, missing sources, and factual errors that would explain what the baseline shows.

Citation tracking is the ongoing work: repeatedly asking a fixed set of questions across platforms — ChatGPT, Perplexity, Google AI Overviews, Gemini — and recording whether the business is mentioned, whether it is cited, and which source is cited. The distinction between mentions and citations, explained in Understanding AI Citations, matters here: a business can be mentioned often without ever being cited, and citation share — the proportion of answers in which a source is cited — is a more meaningful signal than raw mention counts. Tracking must also account for noise: the same prompt can produce different answers on different days, so results are only meaningful in aggregate, over a set of questions and a period of weeks, not from a single check.

A 30/60/90-day cycle gives that aggregation a structure. At 30 days, the question is whether the record itself has improved: new or corrected sources published, inconsistencies resolved, structured data in place. At 60 days, the question is whether retrieval has changed: is the business appearing in more answers, and are the right sources being surfaced? At 90 days, the question is whether the pattern is stable: consistent mentions and citations across platforms, accurate facts in answers, and a record that holds up over repeated checks. None of this guarantees a particular outcome — platforms change their models, competitors act, and measurement itself is imperfect — and any report that promises a specific result at a specific date should be read skeptically. The strategies being measured are set out in Building AI Visibility and Authority.

What this section covers

Questions this section answers

Readers will be able to answer: How do you know whether AI visibility work is having any effect? What is the difference between a mention and a citation, and why does it matter for measurement? Why do single checks mislead, and how many checks are enough? And what should an honest visibility report contain — and what should make you suspicious of one? The foundations of retrieval and citation are covered in How AI Retrieves Information and Understanding AI Citations.

How it connects to the other sections

Measurement closes the loop the other sections open. How Artificial Intelligence Works explains why answers vary; How AI Retrieves Information explains what tracking actually observes; Entities, Knowledge Graphs and Semantic SEO explains the record being audited; Building AI Visibility and Authority explains the work being measured. The full overview is on the Knowledge Center page.

What comes next

This introduction frames the discipline; later articles will work through audit templates, prompt-set design, and reporting formats in practical detail. If you notice an error, or can point to a source that improves an explanation, please use our contact and corrections form.

Latest in this pillar