01Anyone can observe
Rank is observable.
Measured by asking the model from the outside, the way a buyer does — repeatedly, across models, across phrasings.
An AI names three to five brands. That list is the market. ktau puts you on it, at the rank you choose — and tells you up front where your brand needs to be present, and with what content, or that the rank can't be won within your budget.
Rank in that list is scarce, and scarce things have a price. We measure rank as share of recommendation — one attribute, one competitor set, on the live models your buyers actually use.
Measured by asking the model from the outside, the way a buyer does — repeatedly, across models, across phrasings.
One attribute, one competitor set. “Best customer service” and “cheapest in category” are different races, with different winners.
We name the steps to a target rank and price them. Others stop at the score.
Measurement is the diagnosis. It tells you where you are invisible. It does not tell you what a rank costs, or whether you can win it. ktau starts where the dashboards stop.
You set the target. We diagnose, plan the presence it needs, and deliver it — or tell you early that it isn't worth chasing.
Your share of recommendation on every attribute that matters in your category, against every competitor, on the live models. The diagnosis — not the deliverable.
We tell you which ranks are winnable, where your brand needs to be present, and the content it needs there — and which races to walk away from. Unwinnable races are flagged before a dollar is spent.
You set the target; we own it. We produce and place what the plan calls for, then verify the rank on the live model — at a stated confidence, on a stated date.
An answer draws on the live web and on what the model learned in training. A durable rank needs both — and most of the category only works on one.
Pages the model fetches at answer time. This is the half everyone optimises for: it moves in days, and it has to be held — every week, every model release.
What the model learned about your brand in training. Slow to win, and durable once won: the model recommends you with no search at all, and keeps doing so release after release.
An ad vanishes the moment the spend stops. A rank in the model's memory stays — and the next thing you publish builds on it instead of starting over.
That is the difference between a channel you rent and an asset you own.
Tested on brands no model knows and on brands everyone knows, in live AI search and in the model's memory.
We'll tell you where your brand needs to be present and with what content — or that the rank can't be won within your budget. Either answer is worth having before you spend.