How to find it
Select Intent map in the left sidebar, pick a domain, and choose a window of 7, 30 or 90 days. The line under the title states what the screen is built on: N answers to your N tracked questions · N platforms · last N days.Named where buyers decide
The headline is the share of answers that steer buyers toward a purchase (transactional answers) that name your brand, with its base next to it: X of N answers. The sentence underneath gives the same count in words. Once brand roles have been read, it adds how often you were recommended: you are recommended in X of the N that name you, counted across all answers in the window. When there are too few such answers to give a fair share, the figure shows a dash instead of a percentage.What was asked, how AI answered
Two bars split the same answers twice:- What was asked: by the intent of the question (Informational, Commercial, Transactional).
- How AI answered: by what the answer did with it (Informational, Commercial, Transactional).
- N answers are left out: their question has no intent score yet. Unscored questions are counted here, never folded into Informational.
- Only N shops recognised so far, so the buying signal still under-fires. Early in an account’s life, fewer answers read as transactional than will later. Read the Transactional figures with this in mind.
The three figures
Where AI names you, by question type
One row per question type (Definition, How to, Comparison, Best-of list and so on): the share of answers that name you, with the number of questions and answers behind it. Rows are sorted from your strongest type to your weakest.- The type flagged Blind spot is your biggest opening, and the panel’s subtitle spells out its figures.
- A type appears once its questions have been asked enough times. The empty panel tells you how many.
- A row with too few answers shows a dash and an empty bar, not 0%.
Where AI names you, by platform
One row per platform: the share of that platform’s commercial or transactional answers that name you, with the number of answers. Informational answers are left out, so this set is wider than the headline, which counts transactional answers only. A row with too few answers shows a dash.How close your questions sit to a purchase
Your scored questions in four bands, closest to a purchase first: Ready to buy, Comparing, Exploring and Just curious, each with its score range. This panel covers your whole question library, not the selected window. The subtitle says how many of your tracked questions have been scored.Other places to read intent
- A question’s page shows what each platform did with that one question.
- Competitors shows Recommended by AI for your brand and each competitor, over the last 30 days, counted on the answers whose roles have been read.
- The answer viewer shows the treatment, the brand roles and the stage of each answer. See Understand intent.
Timing
After a scan, the bars and the platform panel fill in right away. Brand roles follow a few minutes later; until then, Recommended once named reads not read yet. On the 90-day window, older answers may have no roles. They are left out of the role figures, never counted as zero.What the map does not claim
It does not say why you are absent
It does not say why you are absent
The map shows where you are missing. It cannot see whether you publish anything for those questions, whether a competitor published more, or whether the platform simply answers that way without naming anyone. Open the answers.
A low Transactional count early on usually means few recognised shops
A low Transactional count early on usually means few recognised shops
Check the note under the bars: it tells you how many shops have been recognised so far. The count grows as more of the sites the AI cites are recognised.
A dash is never a zero
A dash is never a zero
A figure is printed only when enough answers stand behind it. “Recommended in 0 of 41” is a finding; a dash or not read yet means there is nothing to read yet, and the two are worded differently on purpose.