What you see each week
Named, not named, or we could not ask.
Someone looking for what you sell asks an assistant and gets a paragraph naming four products. You are in it or you are not, and right now you have no way of knowing which — or how often, or who keeps getting the mention instead.
A run that failed is excluded from every denominator. A vendor outage rendered as 0% is indistinguishable from you actually vanishing, which is the single most misleading thing this category can report.
From each answer
- Whether you were namedPer engine and overall. Failed runs are a dash, never a zero.
- Who got named insteadThe products the engines keep returning, including the ones you have never heard of. They are frequently not who you expected.
- What the answer leaned onYour own pages, a review site, or a community thread. That is where the next piece of work goes.
- The answer itselfStored word for word and reachable from the run, so any number on the dashboard traces back to a sentence a person can read.
Worth checking in any tool in this category: what it shows when an engine was unreachable. Ours shows a dash and says the run was incomplete.
What it does not do yet
It does not give a visibility change a verdict. Answer engines vary week to week in a way Search Console data does not, and we want months of observed variance before claiming a cause. For now you get the measurement and the movement.
By engine, and by measurement
Tracking whether ChatGPT names you
The assistant with the largest audience is also the one your buyers are most likely to have asked before they ever reach the site.
Tracking your visibility in Perplexity
The engine that shows its sources is the one that tells you most about where an answer came from, and therefore where the work belongs.
Google AI Overviews, and why position does not answer it
You can rank third for a query and be absent from the answer sitting above it. Those are two facts and they need two measurements.
Tracking your visibility in Gemini
The engine that taught us the most about how these measurements go wrong, because it fails in ways that look like success.
Share of voice in AI answers
Every answer can name you while each one names six competitors alongside. Visibility would read 100%. That is not the whole story.
What the answer was built on
The percentage tells you where you stand. The citations tell you what the engine was reading when it decided, which is the part work can respond to.
Answer engine optimisation, and what is actually known
A new acronym around an old problem, with a great deal of confident advice attached to systems nobody outside the vendors can see inside.
GEO and SEO: same work, different measurement
The interesting difference is not the tactics. It is that one has a decade of instrumentation behind it and the other has prose.
The rivals the engines keep naming
Every answer that names you also names three or four other companies. Those names are a market map you did not write.
The question set is the measurement
Change the questions and the percentage changes, with nothing having happened to you. This is the part of the method people skip.
Questions
Which answer engines are covered?
ChatGPT, Perplexity, Gemini and Google's AI Overview. Each is a surface a buyer actually opens, which is why the engine rather than the model is the unit of measurement everywhere in the product.
Do you name my product in the question?
Never. A question that names you measures whether the engine has heard of you, not whether it recommends you. Questions come from your own non-branded Search Console queries, and you approve them before anything runs.
Is being cited the same as being named?
No, and they are counted separately. An answer can cite your page as a source while recommending somebody else, which is a different situation and often a more fixable one.
Try it on your own site, free
No card, and the free tier does not expire — the first verdict takes about six weeks to arrive.