Methodology

How we measure AI visibility

Last updated: July 16, 2026

The short version

We ask AI answer engines the same questions your customers ask, then count how often the answer names or cites your business versus competitors. It takes about ten minutes because it runs real queries and investigates real sources — we never return an instant guess.

What "AI visibility" means here

AI visibility is the measured share of relevant AI answers in which your business is named, or your website is cited as a source. When someone asks an AI engine "who's the best [category] in [city]?", either your business shows up in the answer or it doesn't. We turn that into a repeatable number you can track over time.

How a scan works

  1. We build real buyer questions. Using your business category and location, we assemble the kinds of high-intent questions your customers actually ask AI — the "best," "near me," and "who should I hire" style prompts that drive real decisions.
  2. We run them through live AI engines. Each question is sent to AI answer engines through a consistent testing interface. Free audits run ChatGPT and Perplexity; the paid retainer additionally runs Gemini and Claude, for all four major engines.
  3. We check who the answer names. For every response, we detect whether your business — by name or domain — appears, and which competitors, directories, and sources the engine cites instead.
  4. We score it. Your visibility score is the share of answers that named you across all questions and engines. We also break it down by engine and by question so you can see exactly where you're strong and where you're invisible.
  5. We map the citations. We record the sources the AI leaned on — the review sites, directories, and pages it treats as authoritative — because those are the placements you need to win.

That real work is why a scan takes about ten minutes rather than being instant. An instant number would be a guess; a measured one is worth acting on.

From measurement to action

For paid members, each scan feeds a ranked action plan: the specific, prioritized moves most associated with closing your visibility gap — claiming the right listings, strengthening the sources AI cites, and publishing the content that answers the questions buyers ask. Free reports show you the measurement; the ranked plan and the evidence behind it are part of membership.

Evidence before we claim impact

We do not automatically take credit when a score goes up. Completed actions are recorded and then linked to later results, but a change stays marked as "measuring" until the observed movement clears our evidence threshold. Only then do we present it as a proven result. Work that hasn't cleared the threshold is still tracked — it's simply not reported as proven.

Improvement in AI visibility comes from the recommended work actually being implemented. We measure and prioritize; the gains follow the execution.

API benchmark vs. consumer apps

We measure through a consistent API-based testing method, which gives a stable, repeatable benchmark you can trust over time. Results from consumer AI apps can differ from this benchmark — they vary by model version, content freshness, personalization, retrieval, and the user's location. Our goal is a fair, comparable measurement month to month, not a snapshot that changes every time you refresh a chat window.

Limitations we're upfront about

Questions

Want the detail behind a specific number in your report? Email support@getaivisibility.ca.