AI visibility

What the metric covers

AI visibility applies to any surface where a language model assembles a response from retrieved and trained knowledge: ChatGPT, Perplexity, Google AI Overviews and AI Mode, Gemini, Claude, Copilot, and the marketplace assistants Rufus and Sparky. Within an answer, visibility has gradations — mentioned by name, cited as a linked source, or recommended as the specific pick — and each gradation carries different commercial weight.

Most vendor tooling reports AI visibility as a brand aggregate. eCommerce Insights splits it into AI brand visibility and product AI visibility, because for a catalog business the two routinely diverge: marketing copy gets cited while the PDPs that carry revenue do not.

Why it matters for ecommerce

A growing share of D2C consideration is shaped inside AI answers before any site session happens. Even when the purchase lands on the brand's own storefront, the shortlist was often formed by an engine that cited 3–7 sources — and if the brand's SKUs were not among them, the comparison never included them. That shaping is invisible in GA4 until the click, which is why visibility needs its own measurement stream alongside SEO analytics rather than as a replacement for them.

The operational discipline is to track visibility per engine, break it down by query intent, and reconcile it with merchandising priorities — the flagship SKU missing from gift-guide prompts is a different problem than the clearance variant missing from spec-comparison prompts.

Reading the metric: an example

A Shopify brand selling reusable water bottles runs 50 buying-intent prompts weekly across ChatGPT, Perplexity, and Google AI Overviews. The aggregate read might be: brand mentioned in 41% of prompts, cited as a source in 18%, recommended as a specific product in 9% (illustrative figures). Those three numbers tell three stories — the brand is known, sometimes trusted as a source, rarely the engine's first pick — and each drives a different workstream: PR and authority content for mentions, citation-surface work for sources, and product-level PDP fixes for recommendations. The aggregate alone would suggest none of that.

How it relates to neighboring terms

GEO and AEO are the disciplines that move AI visibility; AI visibility score is the composite metric that summarizes it per SKU; LLM visibility is the near-synonym that emphasizes the model; AI discoverability is the upstream precondition — an engine cannot cite what it cannot find and resolve. For the structured-data layer engines parse, see schema.org/Product.

How eCommerce Insights measures it

Per SKU, per engine, on a weekly (Starter) or daily (Growth) refresh: which prompts surfaced the product, which sources were cited alongside it, and how the recommendation was phrased. Each PDP carries a citation score and an agent-readability score, and every gap ships with a concrete fix.

Related terms

• AI brand visibility — the brand-level cut of the same metric.
• Product AI visibility — the SKU-level cut — the unit of revenue.
• AI visibility score — the composite per-SKU summary metric.
• LLM visibility — near-synonym emphasizing the model as the surface.
• Share of voice (AI) — visibility expressed relative to competitors.