Measure AI visibility: one reading does not hold

Who actually comes to read our pages?

Fewer genuine AI crawlers than you would think, because a crawler’s declared user agent can be forged in one line.

Fast Growth Advisors therefore checks every request against the IP ranges the vendors publish. Here is the result on our own servers, over seven days.

Out of 1,734 requests declaring an agent whose vendor publishes its addresses, 1,615 came from elsewhere, or 93 %. Those presenting themselves as Perplexity were Google Cloud machines and a hosting provider, while Perplexity publishes only eight addresses.

Not all these crawlers want the same thing

Counting them together mixes two subjects with different consequences.

Some collect training material, others build a search index, others read a page at the moment a user asks a question.

Direct consequence in robots.txt. Blocking a training collector costs no citations. Blocking an index crawler or a user-triggered reader removes the site from answers.

Many sites block the second kind believing they are blocking the first.

Which pages get picked up, and which never?

It reads page by page, and it is surprising.

Pages with heavy search exposure may never be picked up, while a confidential one may be picked up one time in five. To see it, Fast Growth Advisors cross-checks Search Console with its question probe.

The Observatory is our most picked-up page, at 20.6 % of its impressions, for 34 impressions in total. Our article on the LinkedIn algorithm has 1,043 and is picked up only 1.9 % of the time.

Search volume and engine pick-up do not follow each other.

Some pages are read without ever being cited, like our consulting page, present in search and absent from every generative answer.

Why does a single score mean nothing?

Because an engine that is asked a question does one of three things, and merging them manufactures a false problem. It answers and cites the site. It answers without citing it. Or it does not answer at all.

That third case is an absence of measurement.

Fast Growth Advisors met it while building the barometer question set: on business password managers, two different phrasings produced no written answer at all. This category exists, the market is crowded, and the engine writes nothing. Nobody is invisible there, there is nothing to see.

But the expensive trap lies elsewhere: two pages labelled as cited can be in opposite situations.

A tool measuring once labels both as cited, and it is right both times. What it misses is elsewhere: on the first question Google always takes the same page, on the second it changes and cites the English version three readings out of six.

Citation secured, receiving page not: those are not the same works to undertake.

Does ranking still decide who gets cited?

Largely, yes, but no longer alone.

This is the link everyone looks for between classic search optimisation and AI visibility. Fast Growth Advisors measured it on the fifty barometer questions, comparing the sources Google cites with the organic results of the same query.

Ranking well remains necessary, and is no longer enough.

Three cited sources in ten appear nowhere in the results of the same query in France, and more than one in two in the United Kingdom. A budget that funds ranking alone therefore leaves part of the ground to others.

What do we fix when the measurement is bad?

Everything that blocks, page by page: the Fast Growth Advisors audit tool turns findings into work.

It scores each page across four families and lists, rule by rule, what blocks: a lede too short, a title that does not carry the subject, a FAQ missing from the markup, a paragraph that refers to another without naming it.

On the page you are reading, the audit of 3 September 2026 puts the technical and SEO families above the advised threshold, and GEO below it. We leave that visible rather than remove proof to gain points: the tables add sections without adding questions, which the rule penalises.

We preferred the proof to the points.

These four scores change with every edit to the page. Publishing them frozen in a table would mean publishing a stale measurement from the next revision onwards, which is exactly what we criticise in others.

What correction changes, at the scale of a market

These findings are not specific to us.

The Message-Market Fit Observatory audits the sites of French startups after fundraising, on a grid of clarity and engine readability.

Three detailed audits in four fall below the threshold, and the money raised changes nothing. That last figure counts double: message clarity cannot be bought with a funding round, it has to be worked.

What do we measure on the market questions?

Here is the first half of the inventory Fast Growth Advisors keeps, with for each line the question it answers and the latest result.

Nothing is estimated, everything can be replayed.

What would the dashboard look like?

Like this.

Everything below comes from Fast Growth Advisors measurements, between 1 and 2 September 2026, each block from a different instrument.

Three levels, in the order a director reads them. At the top, what Google answers on our market and whom it gives the floor to instead of us. In the middle, which of our pages it picks up and which it ignores. Only at the bottom, how it happens: the crawlers that visit, the composition of the sources and what the extractors keep.

The order is not decorative.

A dashboard that opens with crawler names makes someone read plumbing when they are trying to find out whether their market knows them.