< GEO FOR SERVICE FIRMS >
Open ChatGPT and ask it for the best firm in your category. It'll name three or four in about four seconds. That's the shortlist now, it gets built before anyone calls you, and most firms have no instrumentation on it at all.
We measure whether you're on it, find what's keeping you off, and do the work to change it.
< WHAT WE KEEP FINDING >
Numbers below are from one real engagement, a retained search firm. The shape repeats across every category we've measured, and it's almost never what the firm expected to be wrong.
Not one of those is a content problem, and none of them show up in a traffic report. Every month they stay true, the assistants get more confident about the firms that are already the answer.
< FROM YOUR SIDE >
None of it reaches your analytics, because the part you're losing happens somewhere you have no instrument. You don't get a report on the calls you weren't invited to.
< WHAT WE DO >
We write the question set your buyers actually use, split three ways — the ones where nobody names you, the ones where your name is already in the question, and the head-to-heads. Then we run it repeatedly, because an assistant answers differently every time and a single prompt tells you nothing. What comes back is every answer in full, every source it read to build that answer, who else got named, and how it talks about you.
This is the part you can’t build in-house, and it’s where every other line gets its orders.
A technical review of your site as the AI crawlers see it, not as Google sees it. Whether your robots.txt is quietly blocking OAI-SearchBot or GPTBot — which alone makes you invisible inside ChatGPT. Whether your content is server-rendered or trapped behind JavaScript that crawlers never run. Whether your Organization and FAQ schema exists in the raw HTML rather than the browser. Plus the conventional SEO audit underneath all of it, because the foundations still have to hold.
Readiness is the input, visibility is the output. Trackers only ever show you the output half.
Assistants assemble your firm from everything written about you, and it contradicts itself more than you’d think — a founding year off by years, service lines you dropped, an office you closed, leadership that changed. We reconcile it across your site, the directories, the press and the profiles. We also check whether another company shares your name and is absorbing your answers. Then we map every source the assistants read on your category and work out which ones are winnable.
The source map decides the plan. Rival marketing pages and independent press need opposite responses.
Comparisons that name real competitors including the ones you’d rather not promote. What you charge and what drives it. The verticals you serve, the emergencies clients call you about, the alternatives to each rival a buyer is weighing you against. We write them and your team approves for accuracy, or we hand your writers the specs if you’d rather run it internally.
Aimed only at questions the measurement showed you absent from. Nothing published on a hunch.
< WHAT SHOWS UP >
The data is yours. We hand over the files, not just the conclusions, because a measurement you can't audit isn't worth much to a firm that sells expertise for a living.
< THE HONEST PART >
A ranking. Assistants are probabilistic and they answer the same question differently on Tuesday than they did on Monday. Anyone selling you a position is selling a number they don't control. We guarantee the work and the measurement.
That publishing alone wins the deal. Getting read puts you on the shortlist. Getting picked still leans on what other people have written about you, and when that's the actual constraint we'll say so rather than sell you pages that can't fix it.
Every assistant at once. We go deepest on ChatGPT because that's where the volume is, and they don't all return the same firms for the same question. The technical and entity work carries everywhere. The measurement is honest about where it's thin.
< BEST FIT >
Selling a product catalog instead? Different engagement, built on feeds and SKUs rather than pages. See the brand side.
< FAIR QUESTIONS >
The first plan is free and there’s nothing to install — send us your domain and your category. After that it’s a monthly retainer scoped to your category, no annual lock, three months notice to leave. We quote once we’ve run your numbers, because pricing it before we know whether you have a plumbing problem or a publishing problem would be guessing.
Your visibility split three ways — discovery, direct and head-to-head — with every measured answer in full text. The firms filling the answers you’re absent from. A source map showing who owns every page the assistants read on your category. A technical readiness pass on your site as the AI crawlers see it. Any factual drift or name collisions we find. And the ranked list of what we’d fix first, with the reasoning attached.
It overlaps at the foundation and diverges fast above it. Traditional SEO is scored on traffic, which rewards pages about how to do the work — those bring in people doing it themselves. Nothing in an SEO audit tells you that OAI-SearchBot is blocked in your robots.txt, that your schema only exists after JavaScript runs, that three directories carry a founding year that’s wrong, or that two thirds of what an assistant reads about your category comes off competitors’ marketing pages. Different question, different instrument.
Your call. We write it and your team approves for accuracy, or we hand your writers the page specs and the question set and you run it internally. Plenty of agencies have better writers than they have measurement, so that split is common.
Parts of it. You’d need to build the question set for your category, run it enough times to get past the noise, log which sources got read on every single response, keep it running as the assistants change underneath you, and reconcile your entity data across the web. Does your team have the tooling for that right now? If it does, genuinely, run it yourself. The measurement is usually the piece nobody can stand up.
Some don’t. But the buyer who’d never admit to it still uses an assistant to build the longlist, sanity-check what a fair price is, and find out who else does what you do. By the time they reach you, that part is settled. You don’t have to believe the channel is big to want to know whether you’re in it, and finding out costs you nothing.
We measured it before we sold it, across five service categories, and we run the same instrument on every prospect before we quote them. We’ll be straight about the limits: assistants are probabilistic, we go deepest on ChatGPT, and anyone promising you a ranking is lying to you. We guarantee the work and the measurement. Not the algorithm.
Send us your domain and the category you compete in. We'll run your buyers' questions, check what the crawlers can read, and show you where you surface, who's getting named instead, and what we'd fix first. In writing, inside 48 hours. No pitch deck, no commitment, and if you're already doing fine we'll tell you that.
Get Your Free Plan