Brand Control: what to do when ChatGPT claims false things about you

It's an uncomfortable moment: you ask ChatGPT about your own company and read a description that's partly outdated, partly simply wrong — a service discontinued long ago, a wrong location, an invented detail. That's not ill will on the model's part, but a consequence of missing facts.

Where the errors come from

AI models answer even when information is missing — they fill the gap with the statistically most likely guess. If there are no clear, citable facts, the model "guesses" on prices, services or reviews. Common causes:

• No llms.txt and no schema markup — the facts aren't unambiguously readable for AI
• Contradictory information across website, LinkedIn, Google profile and directories
• Outdated content still in the index but no longer true
• A robots.txt that locks out AI crawlers — the model only knows old or foreign sources

How to counteract

Brand Control means: give the AI clear, current facts so it no longer has to guess. Concretely that means reconciling and aligning all sources, maintaining llms.txt and schema markup with correct data and placing the key facts clearly on the website. Because the AI picture shifts with every crawl and every model update, this isn't a one-time correction, but an ongoing task.