When an AI Confuses Your Company With a Namesake
Two firms, one name, and an assistant that has to pick. Why the less documented company loses, how retrieval merges the two, and what separates them.
Read moreWhat assistants read and cite about you: your own pages, PDFs, reviews, Reddit, crawlers and llms.txt, and why they repeat outdated or wrong facts.
Two firms, one name, and an assistant that has to pick. Why the less documented company loses, how retrieval merges the two, and what separates them.
Read moreArticle 50 of the AI Act applies from 2 August 2026. Your assistant now marks its text; whether your product pages need a label is another matter.
Read moreYour catalogue has a name for what you sell. The person looking for it describes the job instead, and the assistant answers the description, not the label.
Read moreA model has two ways of knowing anything: what it absorbed while training, and what it fetches while answering. Your launch only exists in the second one.
Read moreYour access log already knows which AI systems have read your site. Here is how to find them, and how to tell a training crawler from a live question.
Read moreBuyers put certifications in their prompts. If yours live only inside a scanned certificate, the assistant will name whoever wrote them out as text.
Read moreA site redesign renames your catalogue URLs and the citations pointing at them quietly stop resolving. What breaks, what survives, and what to check first.
Read moreRetrieval is the difference between an assistant quoting your current catalogue and inventing a plausible version of it. Here is the mechanism, without the jargon.
Read moreAI models state wrong facts about companies with total confidence. Here is why it happens to brands specifically, and a way to check if it is happening to yours.
Read moreWhat has to be true, in writing, on the open web, before ChatGPT, Claude, Gemini or Perplexity puts your company's name in an answer about a technical capability.
Read moreYour tolerance tables and certifications might sit in a PDF no AI can actually read. Here's the exact test to check yours this afternoon.
Read moreOpen your robots.txt and look. Training crawlers and live retrieval fetchers are two different decisions, and most sites block both by accident.
Read moreAn assistant recommending a vendor needs a verdict it did not get from the vendor. Here is why review listings get quoted, and what to check on yours today.
Read moreBefore asking whether AI recommends you, ask whether it knows who you are. On ambiguous names, reference sources, and why you never write your own entry.
Read moreA model never watches your video. It reads the text around it. Which means a video without a clean transcript is, to a machine, almost mute.
Read moreWhy a thread beats a landing page when a model looks for an answer, what you are allowed to do about it, and why anyone selling Reddit placement is selling manipulation.
Read moreA diagnosis before a solution: the usual causes of AI invisibility, ordered by how often they turn out to be the real one, and how to check each yourself.
Read moreThe honest split: the work that genuinely helps, the work that happens nowhere near your website, and the levers that do not exist at all.
Read more