Here's the part that surprises people: invisibility to AI has almost nothing to do with how good, big, or established you are. We've watched a forty-year business get skipped while a competitor founded last year gets named first. The engines aren't rewarding age or reputation. They're rewarding readability — whether they can assemble a confident sentence with your name in it.
An AI answer is built, not retrieved. When a buyer asks "who's the best [what you do] in [where you are]," the model stitches together an answer from the text and structure it can parse across the web. If your site talks about itself in brochure language — "quality craftsmanship, trusted since…" — there's no clean fact for the model to lift. So it reaches for whoever did answer the question plainly, and names them instead.
Three gaps make a business invisible:
- The engine can't tell what you do. You call it "solutions" or "services"; the buyer asked for a specific thing. The model needs the specific thing, in words.
- The engine can't tell where you do it. AI answers are local by default. If your market isn't stated plainly and consistently, you're invisible in the one place you actually compete.
- The engine can't tell why you're the answer. No named proof, no direct answer to the buyer's real question — so nothing distinguishes you from a directory listing.
PRAGMA measures the gap, then closes it.
You can't fix what you can't see. PRAGMA starts by measuring — asking the engines the real buyer questions and recording, per engine, whether you're named. The neutrality is in the method: we publish the ruler and make the result reproducible, so the gap is a fact, not an opinion. Then we shape the answer so the engines can read it — and re-measure to prove the change.
We ask ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews the exact questions your buyers ask, and record — per engine — whether your business is named.
We separate the pillars (answer-engine, parametric, local presence), freeze a baseline, and publish to a registry-verifiable record. We never claim a score predicts revenue — we report what the engines actually say.
Brand 9 Signs is a 40-year business (our own founder's, which is why we could open every number) — and it was invisible to AI. Not because it wasn't good; because the engines couldn't read it. We measured the gap, shaped the answers, and moved the result: Gemini, Perplexity, and Google AI Overviews now name it. Same test before and after, on the record — weak pillars shown too.
Are you the answer?
Find out exactly where the engines can and can't see you — with a check you can run yourself.
Are you the answer? Find out