How we work
From one URL to evidence you can hand over
No integration, no dashboard to learn on day one. You give us a brand; we return the AI's own words about it, measured honestly and proven end to end. Here is the whole path.
A brand name and one website. Optionally, your competitors and target markets.
We find out how AI talks about you, do the optimization ourselves, and hand you a before-and-after report your leadership can read.
Request a diagnostic to see a real report on your own brand →You give us a brand and a website
One URL is enough. No credit card, no integration, no access to your systems. Just tell us who you are and, if you like, who you consider your competitors.
brand Acme Study Abroadurl acme.example.comWe agree on the questions that matter
Together we settle the real, decision-stage questions your buyers ask — in every language they ask them. We optimize for the category questions that shortlist you, not vanity searches for your own name.
“best pathway programs for international students?”“which study-abroad agency is actually reliable?”“is Acme reliable for UK applications?”We sample the engines, continuously
We ask each question many times across the major AI engines, from the right vantages, and store every answer verbatim — with the engine, model, timestamp and whether search fired.
ChatGPTGeminiPerplexityDeepSeekQwenn = 1,240We measure — honestly, with sample size
Raw answers become numbers: how often AI recommends you, who it names instead, and in what tone — by engine and market. Every percentage shows how many real answers it came from.
θ 24%n=320 · ChatGPTSOV #2of 5 named+sentiment neutralWe prove it — verbatim replay
Any number drills down to the AI's own words, replayed like a chat with your brand highlighted and sources aligned. This is the evidence chain a black-box score can't offer.
ChatGPT2026-07-08n=320sha:9f3a…c41d …recommended agencies include Acme Study Abroad and Norbright Education…
★ acme.example.com · icef.comIllustrativeWe deliver the work — and report it
Unlike a pure tool, we do the optimization and plot each action on the θ curve. You receive a diagnostic report as a token-gated PDF — a before-and-after evidence packet you can hand to a board or a client.
Diagnostic.pdfaction shipped → θ ↑
Process artifacts above are illustrative — a real run carries your own brand's numbers.
The first pass is a diagnostic. Then it becomes a standing measurement.
Continue, and the one-off report turns into a continuous instrument: the same questions re-sampled over time, optimization delivered, and each action visible against the curve it moved.
The deliverable
What's in the diagnostic report
- (01)How often AI recommends you, per engine and market — with the evidence count behind each number
- (02)Share of voice against the competitors you named
- (03)Verbatim answer excerpts, with your brand and rivals highlighted
- (04)The sources the engines cite, and your owned-territory share
- (05)Where you're strong, where you're invisible, and what to do first
Before you start
What do you need from us?+
A brand name, one website, and — optionally — a list of competitors and the markets you care about. Nothing more to start.
How long until the first report?+
The initial diagnostic is a same-day-to-few-days effort depending on how many engines, questions and languages we cover. We agree the scope up front.
Is the first diagnostic really free?+
Yes — a scoped diagnostic is free and reviewed before we run it. It's how you see the evidence on your own brand before deciding whether to go further.
What happens after the diagnostic?+
If you continue, the one-off diagnostic becomes a standing measurement: continuous sampling, delivered optimization, and reports that show whether the work moved the numbers.