How to read your first AEOTrack result
The first-result email arrives about 7 minutes after you add your website. What named in 0 of 25 answers means, who was named instead, where the sources are, the one fix, advice-only questions, and what happens on day 2 and every week.
About 7 minutes after you add your website (the median for new accounts), the first check finishes and one email arrives with the numbers and one real answer. This is the moment to judge whether the data is real. Here is how to read that email and the Dashboard behind it, line by line, using a common first result rather than a flattering one.
The subject line is the headline
The subject says the count, not a score. It reads one of three ways:
- "Acme: named in 8% of 25 AI answers" when you were named at all.
- "Acme: 0 of 25 AI answers named you. Here is who they named" when rivals were named.
- "Acme: 0 of 25 AI answers named you. Here is the first fix", or "Here is why", when nobody was.
Zero is a common first reading for a small or new brand. It means no answer in the check named you; it does not mean something broke.
The sample line
"We asked ChatGPT, Gemini, Claude, Perplexity and Grok 5 questions your buyers ask on 29 September, and read 25 answers." That is the whole sample: one live answer per engine per question, web search on. Twenty-five is not a lot, and the email does not pretend it is. Every number below it is "of 25".
An example first result
The homepage shows the same card a new account sees, filled with a composite example rather than one customer's data. Read it top to bottom:
| Line | Example value | What it tells you |
|---|---|---|
| Named in | 0 of 25 answers | How many of the latest answers wrote your name. A link alone does not count. |
| Where any brand was named | 11 of 11 answers named a competitor | Of the answers that named anyone, how many named you. Here, none. |
| Cited as a source | 4 answers, named in none | The engines read your pages and still recommended others. Being read is not the constraint. |
| Per question | "3 rivals named, not you", "Cited, not named", "Nobody named" | Which questions are lost to someone, which are half-won, which are advice-only. |
| One thing to do this week | The roundup page all three engines cite lists your two rivals and not you. Get listed there. | The top task in the plan, which leads with the sites AI reads when it names a rival and not you, ranked by evidence (how often the engines read each one). |
The score itself appears once there are at least three usable answers, and it comes with a ± range: the range to expect between two checks. A first score in single digits with a range wider than the score is normal. A move inside the range next week is normal variation.
Who they named instead
When no answer names you, the Dashboard leads with the brands that were named and how many answers named each. This is the most useful table in the product on day one. Where an answer named any business and not you, this is where you see which one. If the names surprise you, add the businesses among them as competitors (a place or a directory is not a competitor); share of voice and the weekly "who is above you" line are built from that list.
Where they got it from
Next to the brands is the list of domains the engines cited, with the count of answers citing each. The Sources page has the full version across every question. This is where the fix comes from. In one account's 34 questions the engines cited 1,685 sources across 380 domains, and four of the ten most-cited pages were "best tools" roundup lists (what those citations showed). Your top ten will look similar. Read those pages as a buyer and note whether you are on them.
One real answer
The email quotes one answer in full: the question, the engine, and the text, with a caption saying which rival it names. Read it once as the buyer who asked. That is what your prospect saw. Every answer is stored and readable in the app, with its sources, so you can check any number against the text it came from.
Advice-only questions
Some questions get advice, not a shortlist. "How do I choose motor trade insurance" returns a how-to and names nobody. Those are marked advice-only, left out of the score, and counted next to it. Each one has a Replace action. Swap it for a question buyers ask by name and it enters the score at the next check. A question where rivals are named and you are not is not advice-only. It is a real zero, and it stays in.
The "How this score works" panel
Under the score there is a panel with the six facts you need: one answer per engine per question; which engines are on your plan; 25 plus up to 75 when named, 22.5 when an answer only lists your site as a source, otherwise 0; advice-only questions are left out; the ± is the range to expect between two checks; a score older than 30 days is marked stale. The full formula is on the methodology page, and One answer per engine explains why it is built that way.
What happens next
- Day 1. An email with what has run so far and, if there is one, a question a competitor already wins, with the answer. During the trial every question and topic you track is checked again daily while you use it.
- Day 2. The one page to change this week, chosen from the sources the engines read.
- Day 3. Connect Search Console, so the tracked questions come from real searches rather than what we inferred from your site.
- Every week. The weekly results email: score with its range, questions gained and lost, who was named instead or who is above you, and one thing to do before the next check. A Monday digest covers the rest of the team.
- Any day. A manual re-check of a question, to see whether a change was picked up. One re-check within 24 hours of the last check is free; the rest come from your plan's monthly allowance.
If no answer names you yet, Not named by AI yet? has five plain steps that change it. For the whole trial, day by day, see A 7-day trial, day by day.
Frequently asked questions
What does "named in 0 of 25 answers" mean?
Of the 25 answers read (5 questions, one live answer per engine on ChatGPT, Gemini, Claude, Perplexity and Grok, with web search on), none wrote your name. A link to your site alone does not count as being named. Zero is a common first reading for a small or new brand.
What is an advice-only question?
A question that gets advice rather than a shortlist, such as "how do I choose motor trade insurance", which names nobody. Those are marked advice-only, left out of the score and counted next to it, each with a Replace action. A question where rivals are named and you are not is a real zero, not advice-only.
When does the score appear, and what is the ± range?
The score appears once there are at least three usable answers, and it comes with a ± range: the range to expect between two checks. A first score in single digits with a range wider than the score is normal, and a move inside the range next week is normal variation.
Before you sign up
The free audit asks ChatGPT and Perplexity one question your buyers ask and, after your email, shows both full answers, who they named and the pages they read. Free, no account, about a minute, and the question is carried into your account if you start the trial. The trial runs on the Starter plan’s limits for its 7 days, with every question and topic you track checked daily while you use it, and no card.
See what the engines say about you.
Presence, position, citations and sentiment — across ChatGPT, Gemini, Claude, Perplexity and Grok. Free for 7 days, no credit card.
Start your free trial →