SEO Intelligence
See which customer questions bring your website up on Google and Bing, which pages earn clicks and where a competitor ranks ahead of you. It is the first of the four questions: before they can recommend you, they have to find you.
The first question
AI summaries and assistants often draw on pages that search engines have already indexed and show. A page that no search brings to the surface rarely becomes a source: that is why the four questions start from organic search.
One row per lens: what is observed to answer the question and what base it is counted on. The four measures stay separate and do not form an index.
| Question | What is observed | On what base |
|---|---|---|
| SEO · Do they find you? | Rankings on Google and Bing, clicks and impressions from the connected property. | The searches observed and the property’s period. |
| AEO · Do they use your information? | AI summaries, snippets and related questions that answer before the links. | The panel questions that show a direct answer. |
| GEO · Do they recommend you? | Mentions, recommendations and citations in AI assistant answers. | The valid answers of each observation. |
| Trust · Do they trust you? | The tone of reviews, articles and conversations relevant to the brand. | The relevant mentions collected from active sources. |
Rankings over time
Follow your website’s average ranking week by week, separately for each engine. Each point counts only the searches you appear in: a search you dropped out of does not lower the average, it is counted separately.
Search intents
Someone searching “headphones for flying” and someone asking which headphones to take on a long flight have the same need. Panel questions are grouped by intent, and for each one you read the four lenses, each with its own base.
Each question belongs to a single intent. A lens is strong, partial or weak according to declared thresholds; the four cells are never added up, so a high value in one does not make up for a low value in another.
Three sample intents for a fictional brand: for each, the status of the four lenses, strong, partial or weak, with the reason in a few words. The cells are not added up.
| Intent | SEO | AEO | GEO | Trust |
|---|---|---|---|---|
| Choosing headphones for travel | Partial: first page for half the questions | Weak: direct answers cite others | Partial: named, rarely recommended | Strong: positive reviews on comfort |
| Earbuds for running | Weak: no questions on the first page | Weak: no answer cites the website | Weak: named in few answers | Partial: mixed tone on fit |
| Finding out whether the brand is reliable | Strong: the website is on the first page | Strong: answers drawn from the company page | Strong: named almost every time | Strong: mostly positive tone |
Your site
With Search Console and Bing Webmaster Tools connected read-only, you see clicks, impressions and rankings for your pages. Your website’s queries can become panel questions: real search suggests what to ask the assistants.
Four steps: you connect the property, choose a country and a period, bring a query into the question panel and compare the two measures without adding them up.
You authorise Search Console or Bing Webmaster Tools read-only, and can revoke it whenever you like.
You read the queries returned by the source, with clicks, impressions and ranking.
A search that brings clicks becomes a question to ask AI assistants.
Ranking in the results and the assistants’ answer to the same question, never added together.
Who ranks ahead of you
For every panel question you see who holds the first page: competitors, magazines, shops and forums. Knowing who is ahead of you tells you which source to win over, as well as which page of your website to rewrite.
The first six sample results for a search about a fictional brand: ahead of the product page are a magazine, a competitor, a review site, a shop and a forum.
Reading the ranking
An average ranking can improve simply because you dropped out of the difficult searches. That is why every value carries the number of searches behind it, and Google and Bing stay two measures.
The same four observations about search read twice: hastily, without the base, and with the number of searches and the engine behind them.
| Hasty reading | Reading with the base | |
|---|---|---|
| Average position | It moves from six to four: it looks like clear progress. | It improves because two difficult searches dropped out of the average. |
| First page | The website is there, so the search is sorted. | It is there in sixteen searches out of twenty-four: eight remain open. |
| Impressions | More impressions mean more interest in the product. | They also grow when the results page changes. |
| Engines | Google and Bing end up in a single value. | Each engine has its own line and its own searches. |
From search to work
When a question stays off the first page, the Action Center suggests what to change: the page that should answer, the missing content, the source to reach. Then you read the same searches again.
You identify which page of the website should answer that question.
The question and the page text are compared, sentence by sentence.
After the change, the same searches are read again, alongside control questions.
Frequently asked questions
What SEO Intelligence adds to the Google and Bing view, how it works with Bing Webmaster Tools, how a search intent is formed and whether it enters the AI Trust Score.
It groups panel questions by intent and places the other three lenses beside each one. The Google and Bing view remains the place where you read results, AI summaries and website data search by search.
Yes. The connection is read-only, like the Search Console one, and Bing data stays separate from Google data: two properties, two measures.
By grouping panel questions according to the need they express. Each question sits in a single intent, so none is counted twice and the four lenses are read on the same questions.
No. The AI Trust Score is still calculated on AI assistant answers, with its six dimensions. Intents are a way of reading the questions, and the lenses never become a second score.
Related paths
Each feature answers a specific question. From here you move on to related features, to the methodology behind their measures and to help for using them in everyday work.
AI Overviews, AI Mode, featured snippets and related questions: see whether Google’s direct answers use your website and what is missing.
ReadObservation after observation, see whether AI assistants recommend you, which competitor they prefer instead and which sources they lean on.
ReadCollect your brand mentions with date, channel, topic and tone, exclude namesakes and always trace back to the original source.
ReadWhat to change, on which page or source, with what text and in what order: the GEO action plan, checked on the same questions.
Read