Sourced
AI visibility · measured, captured, proven

Your next customer just asked an AI. It answered with three names.

The search bar is being replaced by a conversation that returns a shortlist, not ten blue links. If your business is not one of those names, you were never in the room.

Measured across 7 modelsEN · NL · FRAttribution from your own door
The shift

Small in volume. Large in value.

Most agencies leave the AI channel out of the deck, and on raw volume they are right. But the numbers underneath move the needle. That sliver of traffic quadrupled in a single year, and it converts at 14.2% against 2.8% for classic organic, because whoever walks in from an AI conversation has already had the comparison made for them.

900M
weekly ChatGPT users, asking it questions a search box used to get
4×
growth in AI-referred traffic in a single year, off a small base
5×
higher conversion from an AI-referred visit than classic organic
94%
of B2B buyers already use AI somewhere in the vendor choice
You don't lose visitors to AI. You lose the shortlist. The conversation where three vendors get named now happens before anyone ever sees your site.
The blind spot

The reason nobody can prove their AI work.

Ask any agency to prove the AI work they sold, and it goes quiet. There is a hard reason. Roughly seventy percent of that traffic lands in your analytics as direct traffic. The models report no referral. The customer who discovered you through ChatGPT and converted three weeks later through a search for your name gets credited to Google.

So the whole sector reports on dashboard figures instead of on revenue. That works right up until a finance-accountable person asks what it brought in.

What makes us different

We own the front door. So we can prove it.

Here our path splits from every SEO or GEO agency in the market. They stop at your website. We built the layer behind it.

01 / VISIBILITY
Get named
We make you one of the names the models say, and we measure it per model, per market, over time. Not a hunch, a percentage.
02 / CAPTURE
Catch the demand
Our intake layer answers every inbound contact, day or night, and logs how they found you. It turns the request into an appointment and a file, without anyone having to remember to ask.
03 / PROOF
Prove the result
Because capture lives in your own system, demand from AI is not estimated from broken analytics. It is caught at the door and recorded as your own data.
Visibility, capture and proof in one closed loop. We build the demand, we catch it, and every month we show you which concrete conversations and files came out of it. No classic agency holds that second and third step.
The measurement

Why most AI-visibility reports are noise.

A language model does not give the same answer twice. Research across nearly three thousand prompt runs shows the odds that two identical questions return the same list of brands is smaller than a coin toss. ChatGPT rewrites your question before it answers, and weighs what it knows about the user, including location. Whoever runs a prompt once a week and reads the result as a position is selling you noise.

Thirty to one hundred runs per prompt, per model, never once
Multiple personas and locations per measurement
Visibility as a percentage, never a single position
Measured in EN, NL and FR separately
Bandwidth and spread always reported alongside
Same question set over time, re-run every quarter
ChatGPTGoogle AI OverviewsGoogle AI ModePerplexityGeminiCopilotClaude

The overlap between what ChatGPT cites and what Perplexity cites is about eleven percent. One strategy does not cover every surface, so we measure and optimize per model instead of pretending it is one search engine.

The method

Four layers, ordered by weight of evidence.

We work only with what has provable effect. The order below is the order in which the research finds the strongest link, not the order in which it sells best.

01

Technical readability

Across five hundred million requested pages, OpenAI's crawler ran no JavaScript. What your page builds in the browser does not exist to a model. We make sure the machine can lift clean text off the page. The one layer where the fix is almost binary.

02

Mentions beyond your own site

Of every factor studied, how often your brand is named elsewhere on the web correlates strongest with AI visibility, well above classic backlinks. We map which sources the models cite in your sector and make sure you are correct, current and favorable there.

03

Ranking and freshness

A strong organic position stays one of the best predictors of a citation, and more than three quarters of the most-cited pages were updated in the last thirty days. Diligent SEO and content upkeep, no mystery. Anyone selling this as a separate discipline is billing twice for the same work.

04

Form of the answer

The answer first, the explanation after. Headings that match exactly how someone asks an assistant. Figures, tables and comparisons instead of running marketing text. The fastest effect and the smallest lever, so we treat it as finishing, not strategy.

The fine print, up front

What we won't do, and won't promise.

Google states in its own documentation that no special optimization exists to appear in AI Overviews, and OpenAI states plainly that placement cannot be guaranteed. We agree. We sell no positions, no guaranteed mentions and no exclusivity that nobody can give.

We also buy no reviews, publish no fake forums, and plant no hidden instructions for AI crawlers. That last one is now explicitly against Google's spam policy and has cost companies a large part of their organic traffic this year. What we deliver is a measurement method that survives scrutiny, execution that strengthens your existing findability instead of undermining it, and attribution that comes from your own front door.

No guaranteed positions or bought mentions
No purchased reviews or manufactured sources
No hidden instructions for crawlers
No dashboard theater in place of revenue
How a trajectory runs

From the mirror to a full trajectory.

Every trajectory starts with looking. What follows depends on what you see in the mirror and how much ground there is to win in your sector.

The conversation
The AI Mirror. Twenty minutes, free, live. We ask together the questions your customers ask an assistant, and you watch, on the spot, who gets named.
Week 1
Question set built, competitors defined, measurement configured and linked to your intake.
Week 2
Baseline with bandwidth, technical findings, a source map and a ninety-day action plan, presented.
Weeks 3–6
Technical readability and identity data put in order. First content moves. Intake registers where demand came from, from now on.
Weeks 7–12
Off-site mentions built, content cadence on freshness, first measurable shift in visibility.
Month 4 onward
Continuous optimization and quarterly recalibration. Reporting shifts from visibility to conversations and files.
The next step

Start with the mirror.

Twenty minutes, no cost, no commitment. You see for yourself what an AI assistant answers today when someone asks for what you sell. If you are already in the shortlist, you will know. If you are not, you will know that too, and what it takes to get there.

  • Live scan across every major model
  • Your real competitors, your real market
  • No preparation, no sales deck

Twenty minutes · live · no cost · no commitment