People ask AI what to buy.
Someone gets recommended.
We measure whether it is you. Across 10 AI engines and consumption surfaces, every week, in your market and your language. Then we build the presence that changes the answer.
Illustrative data. Your numbers will be different and measured live. The n=150 refers to the method (prompts × runs), not a real client result.
Search became a conversation. And that conversation has room for three brands.
This is not a future trend, it is the current state. Most companies still optimise for the Google results page, the ecosystem Google itself is replacing.
Google AI Mode went from zero to a billion users in twelve months. Queries doubled every quarter since launch.
The search box stopped being a blank line. It takes text, images, files, video and browser tabs, and answers with interfaces built on the spot.
Google now builds a custom layout per answer: simulations, dashboards, interactive tables. The list of links is officially winding down.
Sources: Google I/O 2026 · Search Central · 19-21 May 2026
10 engines and surfaces. Two routes each.
MODELS, QUERIED BY API
ChatGPT, Claude, Gemini, Grok, DeepSeek, Mistral and Perplexity. Each capable one runs twice: once from training memory, once with live web search. Reporting a single blended number hides which of the two is working, and they are won by completely different work.
SURFACES, AS A PERSON SEES THEM
Google AI Overviews, Google AI Mode and Microsoft Copilot. Not models: what a real person sees inside Google and Bing, queried from your country and in your language. They return different answers per market, and they do not always agree with the models.
Four phases. No black box. All documented.
We do not sell magic. We sell a plan you can read, execute and compare against a result. International platforms measure visibility; what is usually missing is not data, it is someone who knows what to do with it in your market and your sales cycle. We work in that layer, and we are accountable for the number at the end.
Diagnosis
We map your presence across 10 engines and surfaces with real buyer questions from your category, in your language. Where you are, where competitors are, and the exact gap.
Build
Schema, llms.txt, quotable content, work on the pages that decide. We implement directly or brief your team.
Authority
Placement in the sources the models read: technical directories, comparisons, sector publications. Active construction of external citation.
Operation
Weekly measurement, monthly report, continuous adjustment. Direct access over a shared Slack or Teams channel.
Without measurement. With measurement.
Today · where you are
- A buyer asks ChatGPT and gets three brands. None of them is yours.
- Nobody on your team can say what the models answer today.
- Your agency still talks about the results page, while Google itself has said that paradigm is winding down.
- Competitors have been cited for months and nothing was done about it.
In six months · where you can be
- Consistent citation across the engines that matter in your category, measured week after week.
- Share of voice against your real competitors, named, updated weekly.
- Before and after each action, so the effect is measured and not asserted.
- A written plan across four horizons, starting with what pays back fastest.
What people ask before the call.
+Do you work outside Portugal?
Yes. Measurement is localised to your market: the Google and Bing surfaces are queried from your country and in your language, because a user in London does not see what a user in Lisbon sees for the same question. Reporting is in English.
+Is this SEO?
Classical SEO is the substrate and still produces most of the lift. It is not sufficient: a model can rank you well in search and never mention you in an answer, because the answer is assembled from what it knows and what it retrieves, not from a ranked list.
+How long does it take?
The live-search route moves in weeks. The training-memory route moves in months, because it depends on sources that third parties publish and models absorb. Anyone promising ChatGPT presence in thirty days is improvising.
+What do you actually report?
Citation rate, share of voice against your real competitors, average position, tone, and the sources each engine cites. We do not report estimated impressions or potential reach, because those are not measured, they are imagined.
+What is not transferable from your Portuguese work?
The published research, the case studies and the Tier-1 press relationships are Portuguese. The method transfers; the local source landscape has to be rebuilt for your market, and we would rather say that now than discover it together later.
+How much does it cost?
Our published prices are in euros and anchored to the Portuguese market. For work outside Portugal we quote after the free measurement, when we know how many questions, engines and competitors your category actually needs.
Your brand exists.
The AI does not know it.
Thirty minutes, no pitch. Five critical questions from your category, put to 10 engines, with the real answers and a short written read of where you stand. You decide from there.