Methodology

How we measure AI visibility

Short read on the visibility mini: what it counts, how a pass is run, and what it does not claim. Pair with the illustrative sample.

Offline pass

~8 prompts × 3 runs

Live check

3 hero queries, 6 cells

Kept separate

Live ≠ offline SoV

01

Problem

Buyers ask AI who to hire or buy from. If rivals get named and you do not, you lose the shortlist before the site visit.

  • Surfaces include ChatGPT, Perplexity, Gemini, and Google AI Overviews.
  • AI visibility = whether your brand appears in those answers.
  • Classic blue-link rankings alone do not answer that question.
  • The mini shows the gap on a small, honest set of category buyer queries.
02

What we measure

Mention rate: how often your brand and named rivals appear in generated answers for selected buyer prompts.

  • Offline share-of-voice is scored from the offline runs only.
  • Optional live hero rows use ChatGPT Search and Google AI Overview.
  • Live cells are separate paste checks, not folded into the offline total.
  • That split keeps the snapshot readable and comparable.
03

How the mini works

About eight buyer prompts, three offline runs each, then score who gets named.

  • Optional: three hero live spot-checks (six paste cells, skippable if blocked).
  • Surfaces for live: ChatGPT Search + Google AI Overview.
  • Sample page shows illustrative numbers for a fake brand (Northline Advisory).
  • Product process stays ~8×3 offline plus those optional live cells.
04

What it does not prove

The mini is directional. It is not the full paid AEO Diagnostic.

  • Not every query in your category.
  • Engines differ day to day, so one pass is a snapshot, not a guarantee.
  • Does not prove sitemap, llms.txt, or extractability are fixed.
  • Those crawl signals are what the free mini-audit on the AEO page covers.