Brand facts: what they are and why consistency matters

A brand fact is any claim about your company that someone could check and mark right or wrong. Opinions and positioning lines do not qualify. Typical brand facts include:

• Pricing tiers, plan names and what each plan includes
• Founding year, headquarters and ownership
• Product names, integrations and certifications
• Service area, opening hours or headcount, where they apply

A set of canonical brand facts is the single approved version of each one. Writing that set down as a brand fact sheet is the practical first step, and the brand fact sheet template gives you a structure to fill in.

A fact passes through up to four stages before a user reads it.

1. Parametric. A model answering without search uses what it learned before its training cutoff. Nothing you publish today changes that answer until a later model ships.
2. Access. A crawler must be allowed to fetch the page and must actually see the fact. Vercel found that ChatGPT and Claude crawlers fetch JavaScript files but don't execute them, so a price rendered client-side may not exist for them. Googlebot does render JavaScript.
3. Retrieval. The engine decides, per query, whether to search and which pages enter its context. You don't control this.
4. Generation. The model picks which retrieved sentences to use and cite.

You control the first half of the access stage: whether the fact is stated plainly, in the raw HTML, on a page crawlers can reach. That is the part worth getting right.

AI platforms get specific facts wrong about specific brands. A common pattern is a site that contradicts itself: one price on the pricing page, an older one in a blog post, a third in the footer. An engine that retrieves two of those pages has two answers to choose from, and nothing guarantees it picks the current one.

This is an association, not a measured effect. Fixing contradictions does not promise that engines will repeat your facts. It removes one known way for them to repeat the wrong ones, and it is fully in your control.

Severity matters too. A wrong acquisition claim does more damage than a slightly off headcount, which is why one serious AI error can outweigh many small ones. For the full process of finding and fixing conflicts, read the guide to consistent brand facts, or run the free check for contradicting facts on your own site.

Schema markup gets your facts cited

It doesn't, on current evidence. Ahrefs tracked 1,885 pages that added schema against matched controls and found no major uplift on any platform, with AI Overview citations falling 4.6%. Existing markup does no harm, but it is not a lever for AI answers.

An llms.txt file tells engines your official facts

Google's AI optimization guide says Google Search ignores AI text files, and no other engine documents reading them. Whether you can tell AI engines which page is your official source is a separate question with a narrower answer.

Updating the date makes a fact current

A changed dateModified does not change the fact. Keep dates honest and consistent with a date consistency check, and treat them as hygiene, not as a ranking signal.

• Hallucination: an AI answer that states something unsupported or false. See AI hallucinations about brands.
• Canonical brand facts: the single approved version of each fact, used everywhere you publish.
• Brand fact sheet: the internal document that records those canonical facts and where each one appears.
• Server-side rendering: delivering content in the initial HTML so crawlers that don't run JavaScript can read it.

Where should brand facts live on my site? Put each fact on the page a buyer would expect to find it: prices on the pricing page, founding year and headquarters on the about page. State it in plain text in the initial HTML, not only inside an image, a tab loaded by script, or a PDF. Then make every other mention on the site match that wording, or link to that page instead of restating it.

Can I correct a wrong fact that an AI model already learned? Not directly. Answers given without search come from the model's training data, which is frozen at a cutoff. What you can fix is what engines find when they do search: your own pages and the third-party pages that mention you. When engines cite other sites instead of yours, being named accurately on those sites is the route into the answer.

How often should I review my brand facts? Review them whenever a fact changes, such as a price update, a rebrand, a new office or a discontinued product. Search the site for the old value and replace every instance on the same day. A quarterly sweep catches the leftovers in blog posts, help articles and press pages that nobody remembers to update.

Does structured data help with brand facts at all? Google says structured data isn't required for generative AI search, and Ahrefs' matched test found no major citation uplift from adding it. If you already use schema, keep the values in it identical to the visible page text. A schema block that says one price while the page says another is just one more contradiction.