Measured AI visibility · Generative Engine Optimisation
The free check tells you who they recommend. The audit tells you where you rank against the names they actually use, who they cite, what they say about you, why you are absent and what to change. The Tracker tells you whether it moved.
10 questions, a verdict for each, four engines free and six on the audit. No card, no call. See a full example report

Step 1
Before anything is changed, an honest picture of how AI treats your brand: which questions you win, who it names instead, whose sites it trusts, and what it says about you.
Example report · Northwind Coffee
We asked 10 questions your customers ask. You were absent from 6.
“best specialty coffee subscription uk”
Recommended instead: Beanline, Copper Kettle Coffee
“where to buy freshly roasted coffee beans online uk”
Recommended instead: Beanline, Roast & Co
“best coffee beans for espresso machine 2026”
Recommended instead: Copper Kettle Coffee, Beanline
“coffee subscription gift uk”
Recommended instead: Beanline, Copper Kettle Coffee
“ethically sourced coffee roasters uk”
Recommended instead: Beanline
“light roast vs medium roast for v60”
“best single origin coffee beans for filter”
Also recommended: Roast & Co
“best decaf coffee beans uk”
Also recommended: Roast & Co
“how to choose a coffee grinder for pour over”
“northwind coffee review”
Fictional business, illustrative answers. Your report holds this table for your own domain: 10 questions, 4 engines, each asked 5 times. How the probe works
Winning these answers: Beanline (13) · Copper Kettle Coffee (7) · Roast & Co (7)
| # | Name | Questions | Share | ||||
|---|---|---|---|---|---|---|---|
| 1 | Beanline | 7/10 | 31% | ||||
| 2 | Copper Kettle Coffee | 5/10 | 22% | ||||
| 3 | Roast & Co | 4/10 | 17% | ||||
| 4 | Northwind Coffeeyou | 2/10 | 9% | ||||
| 5 | Grind House | 2/10 | 8% |
See your position against the businesses AI actually names for your buyers, read from the answers themselves, not from a list we guessed.
You know who you are really up against in AI, which is rarely who you assumed.
Your site is 3rd of 41 websites the engines cited
Your pages cited: /guides/choosing-a-coffee-grinder (3×), / (2×)
See which websites AI trusts when it answers your buyers, whether yours is one, and which of your pages it used.
You know whose authority to borrow, by getting listed or reviewed there, or to beat.
“Northwind is a small-batch roaster praised for freshness, though pricier than the big brands.”Perplexity
Read how AI describes your brand in its own words, and find the facts it gets wrong, with your site's version beside them.
You manage your reputation where buyers now form it, and correct wrong facts at the source.
Then see the ten questions all of that came from, and correct them if we got your market wrong. The results are yours to trust because the questions are yours to check.
Step 2
The verdicts say what. The readiness half says why, so the money goes on the cause.
Six readiness categories
See what on your site stops AI from citing you: citability, brand, content, technical, schema and platform, scored from measured signals so the same site scores the same twice.
You fix causes, not symptoms.
Content the engines will vouch for
The content category measures experience, expertise, authority and trust, the signals engines use to decide whether a source deserves citing (Google calls this E-E-A-T).
Named people, first-hand detail and real references are what get quoted.
Crawler and platform access
See whether the AI crawlers can read your site at all, platform by platform. We read the HTML your server returns without running JavaScript, because that is what they do.
The cheapest win there is: unblock a crawler and you exist to that engine.
Built to be cited, or actually cited
See whether your problem is structure or authority.
You spend on the right thing: content and schema, or reputation and mentions.
Readiness is the cause. Citation is the effect. The gap tells you where the work is.
Your pages are built to be cited, yet the engines rarely cite them. This is an authority gap, not a structure gap: the fixes are off-site (mentions, reviews, references) more than on-page.
Step 3
Every owner has a different amount of time, budget and technical help. Each route is covered, and every finding comes with the change to make.
A finding, straight out of a real audit
No AggregateRating on any of 50 product pages
Product schema is sound but lacks an aggregateRating property, so Google cannot render star ratings in results. Recommendation: populate it from the Trustpilot API in the product template. The schema block is in the report, ready to paste.
Severity-tagged findings, ready-to-paste JSON-LD and a 30-day plan. Give your developer or agency a prioritised list and a month's schedule.
Progress starts this week without a strategy meeting.
Each section comes with a prompt for Cursor, Claude Code or ChatGPT. Paste it in and it makes the changes, file by file.
Fixes without hiring a developer.
Most popular
Fix packages picked from your own findings, done by us, reviewed by you before anything goes live.
The zero-effort route when time matters more than money.
Step 4
Visibility work is only worth paying for if you can see it pay back. The same questions, asked again, tell you.
Illustrative. Solid dots are audits, keyline dots are scans; readiness holds between audits while visibility and overall move.
Tracker and digest
A monthly re-run on the same questions and an email saying what changed and why: questions gained and lost, rank movement, tone, new claims.
Proof the work paid off, and an early warning when a rival overtakes you.
Real movement, not noise
Every question is asked five times; the score carries its band, and change is measured only on questions both runs asked.
You act on trends, not on a wobble.
Share links and the PDF
Brief your agency, developer or board from the same page, without your login.
Everyone works from the same evidence.
Measured, not guessed: the method is versioned and public, and we run it on ourselves.
Coming next
Two things coming that give your brand a firmer footing in the engines' own knowledge. No dates promised; they are on the roadmap.
A Wikidata entity
An entity record the engines can resolve your brand to, prepared from your site and held for your review before it is published.
The single brand signal that scores highest.
Questions seeded from Search Console
Have the buyer questions start from the queries Google already shows you for, instead of our reading of your homepage.
Measured on real demand from day one.
Pricing
Start free. Pay once when you want to know why, and what to change. Every paid audit passes a QA gate before it reaches you. If it doesn't, you get a full refund.
Step 1 · Am I being seen?
Results in your dashboard in a few minutes.
Steps 1 to 3 · Why not, and what do I do?
One domain. About 25 minutes.
Step 4 · Is it working?
Switch it on for any domain from your dashboard after a check or an audit. Cancel any time.
The check takes a few minutes and costs nothing. The audit takes about 25 and is QA-gated and refund-backed.