Your AI visibility is one question: when your customers ask AI who to hire, does it name you?

AI visibility — also called AI search visibility or AI brand visibility — is whether an assistant names your business when somebody asks it for a recommendation in your market. It is not a position in a list of links: an assistant does not show a list, it names a few companies and moves on. You are either in that answer or you are not.

Because assistants do not answer the same question the same way twice, AI visibility is measured as a share — how often you were named out of all the times the question was asked — and measured separately for each assistant, because they disagree with each other.

First movement usually shows after six to eight weeks, and there is a reason it is not faster. An assistant does not read your site at the moment somebody asks it a question — it answers from what it has already read, and it re-reads the web on its own schedule. A page changed today is a page that has to be found, read and trusted before it can change an answer.

That is why the second measurement in a Deep check sits eight weeks after the first, and why ongoing work measures once a month: a change has to be separated from ordinary drift before anyone calls it a result.

We can’t promise you’ll be named — the assistants don’t publish how they build their lists, and they change what they do without telling anyone. What we can promise is that you’ll see where you stand, what we changed, and whether it moved. First movement usually shows after six to eight weeks. If nothing moves, that will be in the numbers too, and we’ll say so rather than dress up a small swing as progress.

Those are AI visibility trackers: they hand you a dashboard and leave. You read the numbers, work out what they mean, decide what to do, and then do it — or find someone who will. That is four jobs you did not have yesterday, and for a business that is not a marketing department it is usually three jobs too many.

We are not a dashboard you are left to interpret. You get the measuring and the work, and a person who knows your account by name. Not a support queue, not a chatbot — someone you can write to who already knows what we changed last month and why. When a number moves, they will tell you what moved it. When it does not, they will tell you that too.

We also do not try to serve everybody. Sayso is built for businesses whose customers ask an assistant for a recommendation — a service, a tool, a supplier — and the questions, the competitor list and the work all follow from that.

The questions come from your own pages and your market: the way customers ask before they know a single company by name, in your language. Each one goes to each assistant several times over. We record which businesses came back, in what order, and which pages the answer rested on.

What you get is a share per assistant, the questions where you never come up, who was named instead, and every raw answer behind the figures. To track brand visibility in AI over time rather than once, the same questions are put again every month. The long version, including what the method cannot do, is on the How it works page.

ChatGPT, Gemini, Claude and Perplexity. You choose which of them to include, and each is reported on its own — a single blended number would describe none of them.

Google AI Overviews and AI Mode we do not track yet, and we would rather say so than imply otherwise. Measuring them is a different job from asking an assistant a question: it means collecting Google result pages from inside each place you care about, and done carelessly it produces a number we could not stand behind. It is the next thing on our list. When it is real, it will say so here.

Every month on Ongoing work. The questions are put to each assistant more than once each time, because assistants do not answer the same question the same way twice — asking once would record a coin toss.

A month apart is deliberate: measured more often, a share moves inside its own margin of error, and an alert about that is noise rather than news. Questions cannot change in the middle of a measurement: a changed set would look like a changed result, and the two numbers would stop comparing.

The free check covers all four assistants. A deep check starts at $140 — $230 if we carry out the changes as well and measure again eight weeks later. Ongoing work starts at $120 a month, or $200 if we do the monthly work too, with no minimum term.

Two things move the figure: which assistants you want tracked, and whether we do the work or your own team does. ChatGPT and Claude cost several times more to query than Gemini and Perplexity, and that difference is real, not a markup. Tell us your market and the number comes back the same day by email.

If you’ve chosen the package where we carry out the changes, yes — we’ll need access to make them. Nothing goes live without your approval: you see what’s proposed, you say yes, and only then does it happen. Every change is logged in your dashboard with a date and a result. If you’d rather keep access to yourself, take the package where we tell you what to change and your own team does it.

Yes. There’s no minimum term on Ongoing work. A Deep check is exactly what it says — one-off.

Sayso is an AI visibility service: we measure whether assistants name your business when customers ask them for recommendations, and then do the work to change it. Not a dashboard you’re left to interpret — the measuring and the fixing in one place.

It was started by Olha Kryva, an SEO and GEO specialist who has spent the last decade getting businesses found in search — five years agency-side, and five before that running content for an online retailer. She also teaches the SEO module on an internet marketing course at Hillel IT School, including the part about optimising for AI search: Google AI Overviews, ChatGPT, Perplexity. Which is how this started — the questions students asked had no good answers yet, and neither did the tools.

Every client is handled by someone who knows their account, not by whoever picks up the ticket next.