A competitor's comparison page misrepresents us in AI answers. What can I do?

AI answers about your brand come from two places. One is what the model learned in training, frozen at a knowledge cutoff. The other is pages the engine retrieves when it searches the web for a question. A competitor's "Us vs Them" page sits in the second group. If an engine retrieves it for a buyer question, it can quote the page's claims about your pricing, features or limits as fact.

That gives you three moves, in this order:

1. Confirm the page is cited. Don't act on a hunch. Find the answer that carries the wrong claim and check its sources.
2. Publish the correct facts on your own fetchable pages. A specific statement (a price, a plan limit, a supported integration) on a page a crawler can read gives the engine a sentence it can quote instead.
3. Request a correction. The competitor's page is the source. If it changes, the claim stops spreading from it.

What you can't control is which retrieved page the model uses. You can control whether the right fact exists in a readable form, and you can ask for the wrong one to be fixed.

Run the buyer questions that produce the wrong claim on each engine with web search on. Perplexity lists its sources. ChatGPT search shows citations when it searches. Note which URLs appear next to the wrong statement.

Three things can confuse the diagnosis:

• Retrieval is per query. As of October 2026, ChatGPT decides for itself when to search, and OpenAI's help page says it rewrites your question into one or more targeted queries. The same prompt can search one day and not the next.
• No citations can mean no retrieval. If the answer shows no sources, it may come from training data. Then the comparison page may not be the cause at all.
• Cited is not the same as used. OpenAI's API exposes retrieved URLs separately from cited ones, and the retrieved set is larger. A page can shape an answer without being shown.

If the wrong fact appears on several engines, see why the same wrong fact shows up on several AI platforms. For a fuller method, follow how to find which source caused AI to get a fact wrong or the playbook on finding the third-party pages AI engines cite.

Stating a specific fact on a page that can be fetched is the only positive experimental signal for on-page content in the evidence. So write the facts the competitor got wrong as plain sentences: "The Starter plan costs $X a month and includes Y." Put them on your pricing, product and About pages. A fair comparison page of your own can do the same job.

Then make sure engines can actually read those pages:

• Render the facts in the HTML. Vercel's crawler study found ChatGPT's and Claude's crawlers fetch JavaScript files but don't execute them. Facts that only appear after JavaScript runs may be invisible to them.
• Allow the search bots per bot. OpenAI's bots page says OAI-SearchBot controls ChatGPT search and GPTBot controls training, and "each setting is independent of the others". Check PerplexityBot and Claude-SearchBot too.
• Check the firewall. A CDN or firewall can block a crawler that robots.txt allows. The AI Crawler Checker requests a page as each AI crawler and shows which ones get turned away.

This gives engines a correct sentence to find. It doesn't guarantee they'll choose yours over the competitor's.

Most wrong comparison claims are stale, not malicious. A plan changed, a feature shipped, and nobody updated the table. Write to the page owner with:

• the exact sentence that's wrong
• the correct fact, with a link to your public page that states it
• the date the fact changed, if it did

Keep it short and factual. Our playbook on asking third-party sites to fix wrong facts has a template. If they refuse and the claim is false and damaging, talk to your own counsel; that's a legal question, not an AI one.

While you wait, look at the other pages engines cite for the same questions. When engines cite other sites instead of yours, being named accurately on those sites is associated with getting into the answer. Getting listed on pages AI engines already cite covers how.

It's tempting to fight a misleading page with tricks. Don't. Hidden text, content shown only to AI crawlers (cloaking), and instructions planted in pages for a model to follow are all illegitimate. Google's spam policies cover deceptive practices like these, and getting caught can cost you your Google index presence, which also feeds Google's AI features.

Injected content can work in a lab. A June 2026 study of LLM search agents (arXiv 2606.16821) found attack success rates varied widely by model, from none to roughly a third, using a mock search setup rather than the live web. That variation is the risk: the next safety update can turn it off, and your brand is attached to the attempt.

The defensive version is legitimate. If you suspect the competitor's page carries planted instructions, scan third-party pages for injected instructions and document what you find.

How long until AI answers stop repeating the competitor's claim? Nobody can give a date. Answers that search the web can change once the source or your pages change and get recrawled. Answers built from training data only change when the vendor ships a new model, so a fix made this quarter won't reach them. Recheck the same buyer questions monthly and note whether each answer showed citations, so you know which kind you're looking at.

Should I publish my own comparison page in response? It can help if it's accurate and fair. Its job is to put specific, correct facts about you on a page crawlers can fetch. Keep claims about the competitor verifiable and dated, or you create the same problem in reverse. Don't stuff it with keywords or rewrite it in an "AI style"; Google says you don't need to write in a specific way for generative AI search.

Will adding schema markup make engines trust my version? There's no evidence it will. Google says structured data isn't required for generative AI search, and Ahrefs' matched test of 1,885 pages found AI Overviews citations fell 4.6% after schema was added. Existing markup does no harm, but it isn't a lever for AI answers. Plain sentences stating the fact in server-rendered HTML are what the evidence supports.

How do I know whether the fix worked? Rerun the same buyer questions on each engine with search on, and check whether the wrong claim and the competitor's URL still appear. Measure each engine separately, because one can change while another doesn't. The free checker on this site gives a snapshot for up to three buyer questions, which is a quick way to see where you stand before and after.