How can AEOTrack help you?

October 23, 2026

Your AI visibility report (PDF)

What it is

The AI visibility report is the PDF you download from the Reports page. It is one website's AI visibility for the period, laid out to be read by someone who has never opened AEOTrack: a manager, a client, a writer. It is on every plan. The same PDF is what a scheduled client report attaches, rendered from the same data at send time, so a scheduled copy never disagrees with one you download.

To get one: open Reports, pick the website in the header and click Download PDF.


The pages, in order

One set of answers

Every count in the report, on the cover, in the summary, the brand ranking, the question tables and the sources, is taken from the same answers: every answer the engines gave to your tracked questions in the 30 days up to the moment the report is made. A question checked on several days counts once for each day. The cover's period ("30 days to 23 Oct 2026") is that window, and the method page states how many answers it holds. The AI visibility score is the one figure that is not a count of those answers: it is the Dashboard's own score, which reads the newest answer from each engine to each tracked question.

Cover

The logo (AEOTrack's, or yours on a white-label plan), the brand and website, the period, and a strip of three numbers:

  1. X of N AI answers that named you in the 30 days, for example "183 of 520";
  2. your share of voice among the brands you track;
  3. the AI visibility score, 0 to 100, with its ± range.

Under the strip: who the report was prepared for, who prepared it and the date it was generated. Read the ± before comparing the score with an earlier report: a move inside the range is normal variation between two checks, not a change in visibility. See What the ± on a score means.

Contents

A page list. Each section shows its tables, says in a panel what they mean for your brand, and lists up to three things to do next.

Summary

Five lines, each under its own heading:

  • Where you stand: how many of the period's answers named you, as a count and a percentage, your share of voice among the brands tracked, and your AI visibility score.
  • Who leads: the tracked brand named in the most answers, with its Named in rate against yours (or, when you lead, the closest rival's).
  • Widest gap: a question where a tracked rival is named and none of its checks named you.
  • The page to fix: the page on your site the engines already cite most, or, if none is cited yet, the most-cited roundup or source to get onto.
  • Checked next week: what the next weekly check will tell you, such as whether you are now named on the widest-gap question.

Under them are three counts: the questions tracked, the AI answers in the 30 days and the pages cited as sources. If you read one page, read this one.

Brand ranking

Every brand you track, yours included, ranked by Mentions: how many of the period's answers named it. Three columns, each by what it counts:

  • Mentions: the answers that named the brand.
  • Named in: mentions as a share of all N answers in the 30 days.
  • Share of voice: the brand's part of all tracked brands' mentions in those answers, with every question counting the same however many answers it had, so the column adds up to 100.

Named in and share of voice answer different questions, so they differ: Named in says how often answers name a brand, and share of voice how the brands split the mentions between them. The ranking holds only the brands you track; the Share of voice page also lists brands AI names that you do not track, and one that keeps appearing is worth adding as a competitor.

Questions

(Reports made before the October 2026 relabel title this page "Prompts".) Two tables from the same answers:

  • Questions where you are named: up to 10 questions, the ones that name you most first, with Named (how many of the question's checks named you) and Checks (how many checks it had in the 30 days; a check is one answer from one engine).
  • Questions where a rival is named and you are not: up to 10 questions where a tracked rival is named and no check named you, with who was named instead. These are the questions to work on, and each has an opportunity report.

The panel under the tables counts every tracked question, not only the rows shown: how many name you, how many name a rival and never you, and how many name neither.

Sources the AI reads

The pages and sites the engines cited as sources in the same answers, with the total number of citations:

  • By page: the 10 most-cited pages, each with its Type (your site, a rival's site, a roundup or comparison list, a review, news, a guide or another third-party page), its Share of all citations and how many times it was Cited.
  • By site: the 10 most-cited sites, with the same share and count.

A roundup cited often that names your rivals is the page to get onto; the section's next steps name it.

Your pages that get cited

Two tables: the pages on your own site the engines cite most, with how many times each was cited, and up to 10 questions whose answers cite your site. These are the pages to keep current. A question with no cited page of yours means the engines have nothing of yours to lean on for it.

How this report is made

The last page names the engines asked, one live answer per engine per question, and says that every count in the report is taken from the same N answers of the 30 days. It defines Named (the answer mentions you by name; a bare link to your site is not a mention), Cited (the answer lists a page on your site as a source), Named in and share of voice as above. It then says that the AI visibility score is the one figure that is not a count of those answers, what the score is built from, what the ± means, and the scoring version (version 9 since October 2026). AEOTrack's own report ends with a link to the public methodology page. Leave the page in when you send the report on: it is what lets a reader trust the numbers.


White-label (Agency and Enterprise)

On Agency and Enterprise the report carries your branding instead of AEOTrack's: your logo on the cover, your brand colour, your agency name, and no AEOTrack mark if you turn it off. With the mark off, the cover drops AEOTrack's web address, and the method page names no publisher and has no link to AEOTrack's methodology page. Set it once under Settings → Company → White-label branding and every report from every website in the workspace uses it, downloaded or scheduled. The steps are in Share links and white-label reports.


Reading it with a client

  1. Start on the cover: how many answers named the client, its share of voice, and the score with its ± range. If you compare with an earlier report, say whether the score moved outside the range.
  2. Go to the brand ranking: who is ahead, by how many mentions.
  3. Pick one question from the second questions table, where a rival is named and the client is not.
  4. Show the sources: the roundups and sites the engines read, and which of them name the rivals.
  5. End on the summary's page to fix. That is the work for the month.

Frequently asked questions

Is the PDF the same as what the Dashboard shows? It is made from the same stored answers at the moment you download it, but it counts more of them. The report counts every answer of the last 30 days, so its "X of N AI answers that named you" is larger than the Dashboard's count, which reads only the newest answer from each engine to each question. The AI visibility score is the same on both: the report prints the Dashboard's score, its ± range and its scoring version.

Can I get the same numbers in a spreadsheet? The data behind them: the Google Sheets feeds on Pro and up, or the public API. Their windows can differ from the report's 30 days; the API's sources, for example, cover every stored check.

Is there a report for one question? Yes, the opportunity report, which also downloads as a PDF.