AI visibility · in the open

Field Notes

Short, sourced observations from the field: what AI engines said about companies, and what that does and does not show. Each note covers one question or one small test, in the first person, with its sources and its limits.

A note is one observation, not a study. New here? Start with the overview.

Start here

• 001 Ask the same question twice, get two answersAI answers shift across sessions and time. A single check is a snapshot, not the truth.
• 009 AI cites sources that don't check outA citation that looks real isn't proof.
• 014 The ghost in the logged-in answerLogged in, it knew me. Logged out, it had never heard of me.

• 020AI engines don't even agree on what the question isI asked one question. Some engines heard 'recommend a tool,' others heard 'hire a compliance auditor.'
• 019The entity-recognition mountain every new brand starts onYou don't begin at zero. You begin sharing your name with townships, a golf brand, and an insurance agency.
• 018Recommended ≠ cited: two different gamesFrom one audit I built two lists: who the engines recommend, and who they cite. They barely overlapped.
• 017Why AI leans on Reddit — and what to do about it, honestlyThat AI cites Reddit is the finding. Why it does, and what you can honestly do about it, is the useful part.
• 016How to read your own AI-visibility auditYou can check what AI says about you in an afternoon. The hard part isn't running it — it's knowing what you're looking at.
• 015Cited, not ranked: the shift SEO people keep missingRanking earns you a link in a list. Being cited earns you a sentence in the conclusion. Those are not the same job.
• 014The ghost in the logged-in answerI asked an AI what my own project was. Logged in, it knew me. Logged out, it had never heard of me.
• 013Why I never type 'search the web' into an AI visibility testThe instinct to add 'search the web' to the prompt is the thing that breaks the test.
• 012Six engines, six different 'best' agenciesI asked six AI assistants to name the best specialist in my own field. They gave six different answers — no two the same.
• 011Some AI engines cite 14 sources per answer. Others, none.I counted the sources six AI engines show. The spread ran from ~14 an answer down to none at all.
• 010Do AI engines cite the same sources? Barely.In one audit, 83% of cited sources showed up in just one engine. Citation is engine-specific.
• 009AI cites sources that don't check outIn my own audit, AI answers pointed to pages that don't exist. A citation that looks real isn't proof.
• 008AI got a fact about you wrong. The order of operations to fix it.You can't edit the model. You can change what it reads. Here's the sequence: calm, in order.
• 007Why most AEO checklists are wrongThe one-size-fits-all AEO checklist is the wrong mental model. What's portable is the method, not the tactics.
• 006Six engines, five MakefieldsI asked six assistants what 'Makefield' is and got five different companies. None was the one I meant.
• 005Where AI actually looks for your category, and how to be thereTo be in the answer, be where the assistant looks. Find your category's cited sources, then earn a place.
• 004Five engines refused to make something up. One didn't.I fed six assistants a leading premise about a company that doesn't exist. Most refused. One went along.
• 003Ask AI about AEO, and it cites RedditLogging every source the assistants cited about AI visibility, two domains towered over the rest.
• 002What the research says actually moves AI visibilityA Princeton study proved content changes lift AI visibility (up to 40%). The catch is the lesson.
• 001Ask the same question twice, get two answersAI answers shift across sessions and time. A single check is a snapshot, not the truth.