Classic search
You show up in the results. We count impressions and clicks of your pages in Google Search.
1Be one.
Footnote measures what Google and the AI assistants actually do with your pages, keeps every number apart, and tells you what to publish next so you become the source they cite.
Most AI visibility tools run synthetic prompts against the models and roll the answers into a single score.
Ask an AI the same question again and again, and the same list of brands comes back less than once in a hundred tries.1 A score built on that moves when nothing moved.
We start from what was observed: your pages, your queries, your visitors.
Google search, Google’s AI answers, visits from AI assistants. We keep them apart because they record different moments of discovery.
Which AI crawlers request which pages, and what your site answers them.4 It shows what gets fetched, not what gets cited, so it never becomes a fourth score.
Connect Search Console and Analytics. We work out what you do and who searches for it, match the questions those people ask AI against the queries you already appear for, and turn every gap into a brief your team, or ours, can produce.
Has savings sitting in a bank account and wants to start investing without a costly mistake. Asks ChatGPT before she asks anyone.
Illustrative persona and numbers. A prompt with no matching queries stays flagged until data backs it.
Four onboarding steps, done once with our team. You connect, we do the reading. Then the loop starts: produce, measure, decide what comes next.
Your domain and the markets you sell in.
Search Console and Analytics 4, read-only. Cloudflare if you use it.
From your queries and pages we work out what you do and who searches for it.
Their prompts, matched to your real queries.
Hubs, articles, PR, podcasts. Same three tracks.
We build each hub around one precise research question. It becomes a reference point that organises articles, PR, podcasts and dedicated pages, and links them to everything your brand already owns.
Then we measure what each hub earns on the three tracks and decide what to produce or fix next, even when the numbers are small.
Google's own guidance points the same way: original, useful, accessible content is what its AI features draw on.5 Hubs are our design choice for doing that at scale. We are validating them in the open.

*Footnote was born from our clients' needs: problems we solved for them as tools first, then turned into a product.
We run prompt tests as research, to find gaps. We never report them as your performance, because the answers change from one run to the next.1
Your website, read-only access to Search Console and Google Analytics 4, and Cloudflare if you use it. We work out the rest and review it with you in one call.
Yes. Small observed numbers tell you more than large simulated ones. We show them as they are and name the page that earned each one.
Fylle's editorial team, yours, or both. Every brief arrives with the persona, the prompt and the queries behind it.
Every numbered chip above, like this one1, marks a claim. Here is the source behind each number.
Official sources back the measures. The hub method is our own design choice, validated on live client projects.