What to Do When AI Gets Your Company Description Wrong: Fixing How ChatGPT Describes You with a Brand Facts Page and Organization Schema
Why do ChatGPT or Perplexity get my company's services, pricing, or team wrong?
AI models have a training data cutoff date, and they prioritize the most commonly found information online rather than official sources. If your official information isn't well indexed, lacks structured markup, or outdated information is more prevalent online, AI will provide inaccurate answers.
ChatGPT's training data is typically several months behind its release date; while Perplexity crawls the web in real time, it still relies on web visibility and authority scoring. Small businesses are especially prone to misrepresentation due to scattered information or lack of structured markup.
What should a brand facts page include?
A brand facts page should include: the company's legal name, founding year, headquarters address, core business description, list of main services/products, current leadership team, contact information, and official website link. Each piece of information should correspond to Organization Schema and be presented in a concise, structured format that's easy for AI to extract directly.
This page is typically placed at /about or /brand-facts, using a consistent, easy-to-parse format (such as tables or lists). Search engines like Google and Bing, as well as AI assistants, prioritize reading this type of factual information from structured data.
How should Organization schema be marked up so AI adopts my official statements?
Embed Organization schema in JSON-LD format within the website's HTML head, including name, url, logo, description, foundingDate, address, and contactPoint. Ensure the markup information is accurate, complete, and consistent with the page text. Use Google's Schema Validator to verify the markup is correct.
Example: "@type": "Organization", "name": "[Official Legal Name]", "url": "[Official Website]", "sameAs": [official accounts on LinkedIn, Facebook, etc.]. Schema.org's sameAs attribute helps AI link to your official identity.
After making corrections, how long will it take for AI's answers to update?
Static models like ChatGPT don't update automatically and must wait for the next training version (typically several months to a year). Real-time AI like Perplexity and Bing Chat usually re-crawl and update answers within 24-72 hours, though updates are delayed if the website uses robots.txt or noindex.
Ways to accelerate updates: submit a URL inspection through Google Search Console, use a schema.org sitemap, and publish updates through official press releases or PR platforms. Continuous SEO optimization can increase a webpage's weight in search results, thereby speeding up AI's adoption of updated information.
My small company doesn't have a Wikipedia page—can I still get AI to describe it correctly?
Yes. Building a complete Organization schema, a brand facts page, a structured About page, and publishing official information on high-authority sites (such as industry directories, business registries, and press release platforms) can compensate for not having a Wikipedia page. Consistent verification across multiple sources can improve AI's confidence in the information.
Priority strategies: verifying your Google Business Profile, pages for industry awards or certifications, media coverage links, and a fully completed LinkedIn company page. AI determines factual accuracy through consistent information across multiple trusted sources, rather than relying on a single Wikipedia page.
Hogan