It is an agent for search and AI visibility. You give it a web address; it reads the site, measures where the project stands in Google and in the assistants' answers, puts the niche and its rivals on one map, proposes a strategy for the month, and then does the work in the order the project can win it: the pages, the checks, the briefs. Led by the agent, not learned from tabs.

Two of your rivals are named on “crm for a five person team” and you are not. I would start there.
The shape of it, drawn at a size the figures can be read. Placeholders, not a customer's numbers.
Two of your rivals are named on “crm for a five person team” and you are not. It has 3,600 searches a month and nothing on your site answers it. I would start there.
Answer first, every figure traced to a source, ready to open in about ten minutes.
Give it a web address. The site is read the way an assistant's crawler reads it, free, in about a minute.
Google's answers, the assistants and the rivals are measured, and the niche is put on one map.
The agent proposes the strategy for the month and the first move, with the reason.
You agree, correct or ask. It does the work and tells you where the result went.
A visibility score tells you that you are losing. It does not tell you which question you are losing, who is being named instead, what would have to exist for that to change, or who writes it. Every tool in this category stops at the number, and the number is the cheapest part.
The agent starts where the number stops. It reads everything the setup collected, proposes one move by name with the reason, and does it: the missing page, the draft, the checked figure, the brief for the developer. Each of those is a step somebody would otherwise do by hand, and each is where the month actually goes.
There is evidence about which of that work pays. According to Princeton and Georgia Tech researchers, whose KDD paper tested 9 content changes across 10,000 queries, the right ones "can boost visibility by up to 40% in generative engine responses". Their own table separates the levers: quotations moved visibility 41 per cent, concrete statistics 40 per cent, named sources 28 per cent.
Search demand, the answers, and the category's signals. Search demand is classic SEO: phrases people type into Google, grouped into pages by a clustering model our own team built. The answers are AI visibility: the questions where ChatGPT, Perplexity, Gemini or Google's AI answers name a rival and not you. The signals are the predictive lane: what the accounts your buyers follow are publishing this week, written into before it reaches a search box.
All three feed one queue, and the strategy for the month says how much each lane counts. The agent proposes the split from what it measured; you can move it whenever you like.
The site check is free and unlimited. It reads your pages the way the crawlers behind assistant answers read them, and reports what is stopping you being quoted: pages assembled in the browser that hand a crawler an empty shell, pages too thin to answer their own question, missing dates, blocked crawlers.
It says how much of your site it read and how much it did not, because a finding about 8 pages on a site of 143 is a claim about the other 135, and hiding that would be the first dishonest thing we did.
The measurement is, and so is the proposal. The writing is not published anywhere without you: a draft is saved as a document you own, with anything we could not verify listed separately for you to decide about.
Nothing is ever posted to your site on your behalf. When you publish, you say so, and from then on the effect of that page is tracked against what it was written for — starting about 14 days later, because nothing finds and weighs a new page faster than that.
Last updated 3 September 2026
Start with the free check. Your web address and nothing else.