A customer asks ChatGPT about your price and gets the rate from two years ago. A guest reads in Google AI Overviews that they can cancel free of charge up to 24 hours before arrival, although your policy is 14 days. AI misinformation about brands very often concerns prices and policies, exactly the details customers base their decisions on.
The good news: in most cases the AI invents nothing. It repeats what is written somewhere on the web. This guide shows where the wrong details come from and which platforms help you correct them.
- AI misinformation about prices and policies usually comes from outdated sources: old price pages, PDFs, portals and comparison sites.
- What helps most is your own website, Google's channels (Search Console, Merchant Center, Business Profile) and the portals the AI copies from.
- Reporting to ChatGPT or Google is worth doing but does not last.
- Fact checking platforms like Profound FactCheck, Scrunch, Bluefish or faktado.ai detect false AI claims and show the source.
Why AI so often gets prices and policies wrong
A business's name and address rarely change. Prices, cancellation policies, return periods, delivery terms and house rules change all the time. Every change leaves old versions online, and those are exactly what an AI reads too.
- Old pages stay online: the 2024 price list is still reachable under an old address or linked as a PDF.
- Portals copy details only once: a booking or comparison portal stored your terms when the listing was created and never updated them.
- Several versions on your own website: the terms say 14 days, the FAQ says 30 days. The AI cannot know which applies.
- Prices without context: “from 89 EUR” without season, room or service is often repeated as a fixed price.
AI systems with web search read several pages for a question and summarise them. If the sources contradict each other, the AI often picks the one that is easiest to read or ranks best in search. Why that is explained in the article on AI hallucinations.
Which platforms can help correct brand misinformation about prices and policies in AI outputs?
The key distinction is between AI visibility and AI fact checking. Many GEO tools only show how often your brand appears in AI answers. For wrong prices and policies you need platforms that detect false claims, compare them with your own details and show the source of the error. Since 2026 there are dedicated tools for this.
Platforms that detect false AI claims
| Platform | What it does | For whom |
|---|---|---|
| Profound FactCheck | Since July 2026. Compares claims by ChatGPT, Perplexity, Gemini and others with a knowledge base built from your website and documents, shows the sources behind wrong answers and triggers corrections through its own agents. | Large companies, included in enterprise plans, price on request |
| Scrunch Knowledge Studio | Manages approved brand information, detects false AI claims and conflicts on your own website with a fact check agent and serves AI systems a version of your pages prepared for them. | Enterprise customers, currently in beta |
| Bluefish AI Accuracy | Since May 2026. Breaks AI answers into individual claims, checks them against the brand's details and scores errors by severity, including for shopping assistants. | Large corporations and heavily regulated industries |
| faktado.ai | Reads the facts on your website, checks the answers of ChatGPT and Google as well as portals, maps and directories against them, shows the source of every error and prepares the correction request, with a re-check. | Small and mid-size businesses in the DACH region, free check, plans from 49 EUR per month |
Pure visibility tools like Peec AI, Otterly.AI or Reaudit show whether and how your brand appears in AI answers and which sources are behind it. They do not describe a comparison with your approved prices and policies. More in the tools comparison.
Sources where you fix the errors
Even the best tool does not correct the AI directly. What lasts is correcting the places the AI takes its details from:
| Platform | What you do there | Effect on AI answers |
|---|---|---|
| Your own website | One current page per topic: prices, cancellation, returns, shipping, house rules. Remove or redirect old versions. | High. It is the template for every other source. |
| Google Search Console | For online shops: add shipping and return policies directly. | High for Google. According to Google these settings take precedence over the markup on your website. |
| Google Merchant Center | Maintain products with price and availability, plus shipping and return settings. | High for Google Shopping and product answers. |
| Google Business Profile | Keep services and products with prices, opening hours and attributes up to date. | High for local questions in Google, Maps and AI Mode. |
| Portals of your industry | Change your terms in the partner area, for example the extranet of a booking portal, or contact the operator. | Medium to high, depending on how often the AI cites the portal. |
| Feedback tools of AI providers | Rate and report a wrong answer in ChatGPT, Google or Gemini. | Low and without any promise. Useful as an extra, not as the fix. |
| Product feeds to AI providers | For merchants: OpenAI accepts product feeds with price and availability. Instant Checkout in ChatGPT is currently limited to approved partners. | For product questions in ChatGPT, still being built out. |
How to report a wrong answer on each AI platform is explained step by step in AI answers wrong? How to correct them.
Publish prices and policies so AI reads them correctly
The most important platform remains your own website. What is stated clearly there wins over time in other sources and in AI answers too.
- One page per policy: cancellation, returns, shipping, payment, pets or house rules each in exactly one place. Other pages link there instead of repeating the text.
- As text, not as an image or PDF: prices and terms belong in the visible page text. A PDF can be offered in addition.
- With context and date: “Double room from 129 EUR per night in low season, as of March 2026” is clear. “from 89 EUR” is not.
- Remove old versions: delete old price lists and earlier versions of your terms or redirect them to the current page.
- Structured data that matches the text: for online shops Google supports return policies as structured data, including once for the whole business instead of for every product. Note: according to Google, settings in Search Console or Merchant Center take precedence over the markup.
According to Google, AI Overviews and AI Mode need no special AI markup. What matters is that the details are clear and the same everywhere.
Examples by industry: where the wrong details usually come from
Which platform matters most depends on your industry. The businesses in these examples are made up, the patterns we see every day.
- Hotel: the AI says free cancellation up to 24 hours before arrival because a booking portal shows an old rate. Fix: align the terms in the portal's extranet and on your own website, close old rates.
- Online shop: ChatGPT says 14 days for returns instead of 30. The source is an old FAQ page. Fix: one returns page, redirect the old FAQ, add the return policy in Search Console or Merchant Center.
- Service provider: the AI quotes an hourly rate from 2022 because a business directory still shows it. Fix: claim and correct the listing, keep prices in your Google Business Profile up to date.
- Restaurant: the AI quotes a lunch menu at an old price from a review site. Fix: put the menu on your website as text instead of a photo, update the portal listing.
In every case: first check which source the AI cites. Perplexity shows sources as footnotes, ChatGPT and Google usually link them below the answer.
Keep AI misinformation in view for good
Prices and policies change, and AI systems read the web anew every day. A one-off correction is therefore not enough. A fixed routine works well: after every price or policy change, and once a month on top, ask ChatGPT and Google the same questions and check the sources they cite.
Pure visibility tools measure how often your brand appears in AI answers. Whether the details are right is checked by fact checking platforms like the ones above. The big providers mainly target corporations with their own knowledge base. faktado.ai is built for smaller businesses: it compares the answers of ChatGPT and Google and the listings across the web with your website, shows the source behind every error and prepares the correction. The free check is the place to start.
maxonline, the SEO and GEO agency behind faktado.ai, takes care of it for you.
FAQ
Can I get ChatGPT to state my prices correctly?
Not directly. A correction in the chat only applies to your own conversation, and reporting an answer comes with no promise. What lasts is correcting the sources ChatGPT reads its details from. Merchants can also send OpenAI a product feed with prices.
Which platform picks up corrections fastest?
The channels you maintain yourself work fastest: your website, your Google Business Profile and, for online shops, Search Console and Merchant Center. With third-party portals it depends on whether you can edit the listing yourself. No AI provider gives a fixed timeframe.
Does structured data help against wrong AI answers?
It helps if it matches the visible text. Google supports return policies as structured data, for example. According to Google no special AI markup is needed. More important is that every detail appears only once and clearly on your website.
What is the difference between AI monitoring and correction?
AI monitoring measures whether and how often your brand appears in AI answers. Fact checking also compares the claims with your own details. Correction finally means fixing the wrong details at the source. For wrong prices and policies you need fact checking and access to the sources that cause the errors.
Harald FischlManaging Director, maxonlineHarald is Managing Director of maxonline, an SEO and GEO agency from Lower Austria focused on hospitality. He works every day on how businesses are presented correctly in search engines and AI answers. More about Harald