Find out whether AI answers cite you — or your competitors
We measure how often answer engines name your brand across a fixed set of your real commercial queries, show you who gets cited instead, and tell you which gaps are worth closing.
Measurement first, because most of this field is guesswork
A large amount of what is currently sold as AI optimisation is unmeasured. Vendors describe tactics without ever establishing whether the brand was cited before or after, which makes the work unfalsifiable and the invoice unarguable.
We start with a baseline. A fixed set of prompts built from your actual commercial queries, run repeatedly, producing a citation rate you can compare against later. Any single AI answer is noise — the same prompt returns different text each time — but the rate across a stable prompt set moves in ways you can act on.
What the audit tells you
- Whether you have a problem at all. Plenty of brands assume they are losing to AI answers when their clicks fell for ordinary reasons. We check impressions against clicks per query before anything else.
- Who is cited instead. Usually a mix of publishers, aggregators and one or two competitors — and the pattern in what they publish is the actionable part.
- Whether you are machine-legible. Answer engines need to resolve your brand to an entity. Missing
Organizationmarkup, no author entities and nosameAslinks make that harder than it needs to be. - What is worth doing. Split into durable work and provisional tactics, clearly labelled, so you are not rebuilding your site around something that may not survive the year.
What we will not tell you
That we can guarantee citations. That there is a submission process. That we have a relationship with any answer engine. None of that is true of anyone, and a vendor claiming otherwise is the clearest signal available that you should keep looking.
Much of what improves AI visibility is work you should be doing anyway — clear content, credible authorship, correct structured data. We would rather sell you that honestly than dress it up as a new discipline.
Questions we get asked
Can you guarantee we get cited in AI answers? +
No. There is no submission process, no ranking dial and no vendor with privileged access. What we can do is measure your current citation rate, identify what the cited sources have that you do not, and close those gaps.
Is this just SEO with a new name? +
Substantially, yes — and we would rather say so. Answer engines draw heavily on the same signals as search: clear content, credible entities, structured data, and links. The genuinely new parts are measurement and being explicitly citable rather than merely rankable.
How do you measure something that changes every time? +
A fixed prompt set, run repeatedly, reported as a rate rather than a binary. Any single answer is noise; the citation rate across fifty prompts over time is a usable signal.
Should we be worried about losing traffic to AI Overviews? +
Check before worrying. In Search Console, compare impressions against clicks by query — if impressions held while clicks fell, an answer is absorbing the click. If both fell, this is not your problem and a traffic drop audit is the right service.
This area moves fast. Will the advice be stale in six months? +
The measurement approach will hold. Specific tactics may not, which is why the deliverable separates durable work — entity clarity, structured data, genuinely citable content — from tactics we flag as provisional.
Read before you hire
- Getting Cited in AI Answers: What Actually Carries OverWhich existing SEO work transfers, and which does not.
- Schema Types That Actually Earn Rich ResultsEntity markup is how answer engines work out who you are.
- Why Did My Organic Traffic Drop?How to tell an AI-absorbed click from a lost ranking.
Start with the measurement
Send us your domain and the ten queries that matter most commercially. We will come back with your current citation rate and who is being cited in your place.
Get a visibility baseline →