The client conversation you'll have when their AI citation says something wrong

This will happen to every agency doing AI visibility work: a client calls because ChatGPT described their business incorrectly. Wrong location, wrong specialism, outdated description, or — worst case — confused them with a competitor. Here's how to handle it, because the fix is less obvious than it looks. **Why it happens** AI systems build their knowledge of a business from a combination of sources: your website, directory listings, press coverage, Wikipedia if you have an entry, and general web content. If those sources are inconsistent or outdated, the AI synthesises an answer from the inconsistency. The result is often a blend of accurate and inaccurate information that's harder to correct than a purely wrong answer. **The fix is not contacting the AI company** There's no "submit a correction" form for ChatGPT or Perplexity. The way to correct an AI citation is to correct the underlying sources the AI is drawing from. That means: update your website content, fix directory inconsistencies, update your Google Business Profile, and — if the wrong information is coming from a specific source like an old press article — contact that source directly. **The timeline** AI systems re-index content at different rates. Google AI Overviews update relatively quickly (days to weeks). ChatGPT's training data updates on a longer cycle (months). Perplexity, which uses live retrieval, updates fastest. Set client expectations accordingly: fixing the underlying sources is the right move, but the AI citation may take weeks or months to reflect the correction. **Where AI Verified helps** The Passport gives you an authoritative, machine-readable identity record that AI systems can query directly. It doesn't override all other sources, but it does give the AI a single verified source to anchor to. For clients with persistent citation problems, getting the Passport data right is a good starting point.