Why this exists
The room you're not in
Somewhere right now, an AI is describing your company, and no one asked you. There's a version of you you never wrote, and a buyer is reading it first. Makefield exists to read that version, measure it, and publish what's true about how it works.
The version you never wrote
Someone is asking about you right now
A buyer opens an AI assistant and asks which company to trust. In a second or two they get an answer — confident, fluent, and built from whatever the model gathered about you from across the web. You weren't asked. You didn't write it. You may not even know what it said.
That answer is a version of your company you never authored. It can be flattering, or wrong, or it can quietly recommend someone else. And it is increasingly the first version a buyer meets — before your site, before your deck, before a conversation. The unsettling part isn't that AI talks about you. It's that it does so from the outside, from fragments, with total confidence, in a room you're not standing in.
The fog
Lights in the fog
Think of it as ships in fog. A model trying to describe your company is a ship reading faint lights through the mist — a mention on Reddit, an old profile, a competitor's comparison page, a press line from two years ago. It steers by whatever it can see. If your lights are few and scattered, it sketches you from guesswork. If they're clear and consistent, it sees you as you are.
So visibility here isn't a billboard you buy. It emerges — from many real signals across the places models learn from, lining up to say the same true thing about you. The work isn't a louder horn — it's more lights in the fog, so the ships see you clearly, and often.
Why research, not a pitch
The field is loud and short on proof
This corner of the work is full of confident assertion and thin on measurement. Plenty of advice; very little anyone has actually checked. That's the gap Makefield was built for: to treat "what the AI says about you" as something you can measure the same way twice — score it, date it, back it with evidence — instead of something you have a feeling about.
And the choice underneath all of it: I publish to be cited, not to sell. Others sell courses and dashboards; the bet here is that the most durable thing in a blurry new field is honest, measured, openly-shared knowledge: the kind a model, or a person, comes to trust and repeat. So the method is in the open, and so are the findings.
Says who?
Why the measurement has to come from outside
A measurement is only as good as the distance between the person measuring and the result. Makefield sells no fix and gains nothing if an answer gets better or worse, and it writes the correct answer down before it asks. That makes a result something another person can check, not something they have to believe. You can look over your own car, but someone else does the safety inspection.
The honest part
Small, measured, and said plainly
This is early, and I won't pretend otherwise. What is published so far is small: single observations, each with its sources and its limits. When a planned test finds nothing, that will be published too. A finding you can check beats a claim you have to believe, every time.
That's the whole of it. There's a version of your company being written without you, in a room you're not in. Makefield is a research project for reading it, measuring it honestly, and putting what's true about how it works back into the open — so the next answer is a little clearer than the last.
Read the answer you didn't write.
It's free to read and cite. There's nothing to buy. Start with the overview, follow the open course, or read the field notes — the proof behind the writing is the point.