Ask an AI assistant "how much does [your service] cost?" and one of three things usually happens. It says pricing isn't publicly available and suggests contacting the company. It quotes a figure from a three-year-old forum thread or review site. Or it gives a confident estimate that nobody at your company would recognise. All three cost you: the first sends price-sensitive buyers to competitors who do publish numbers, and the other two set the wrong expectations before a prospect ever reaches your site.
AI tools can only repeat pricing they can find. When your site says "Contact us for a quote" and nothing else, the gap gets filled by other sources, or by nothing at all.
This guide explains why AI answers struggle with pricing, the options for publishing price information even when every job is custom, how to structure pricing pages and markup so figures are read correctly, and how to check what AI tools currently say about your prices.
Key Takeaways
- AI answers report pricing as "unavailable", or guess, when your own pages do not state it clearly.
- You do not have to publish a fixed price list: ranges, "from" prices, typical project examples and cost factors all give AI tools accurate material.
- Put prices in HTML text with currency, billing period, tax status and conditions, not only in images, PDFs or calculators.
- For products, Product and Offer structured data should match the visible price exactly.
- Audit AI answers for price questions regularly and correct outdated figures at the source.
Why AI answers get pricing wrong
Price questions are among the most common sub-questions in buying prompts. Semrush's August 2026 guide to query fan-out lists price, refunds and comparisons among the subqueries commercial pages should answer directly. When an AI system searches for your price and finds nothing on your site, it has a few options:
- Say it is not available, which is accurate but unhelpful, and often followed by competitors who do publish prices.
- Use third-party figures from review sites, directories, forums or old articles, which may be outdated or wrong.
- Infer from similar businesses, producing a generic estimate attached to your name.
OpenAI's own help page for ChatGPT search warns that results can be incomplete, outdated or incorrect and encourages users to check the cited sources. The best way to make those sources accurate is to be one of them.
Pricing disclosure options
| Option | Example wording | Best for | Trade-off |
|---|---|---|---|
| Fixed price list | "Standard plan: £29 per user per month, billed annually, excl. VAT" | SaaS, products, packaged services | Competitors see your prices |
| "From" price | "Boiler servicing from £85" | Services with a common entry level | Must reflect a real, available price |
| Typical range | "Most kitchen fittings cost between £X and £Y" | Custom work with predictable bands | Outliers need explaining |
| Worked examples | "A 3-bedroom house in Leeds, completed in 2026, cost £X" | Projects, B2B services | Needs regular updating |
| Cost factors only | "Price depends on size, materials and access" | Highly variable work | Weakest for AI answers; add at least a range |
| Calculator | Interactive estimator | Many variables | Add example outputs in text, as calculators are often JavaScript-only |
(Figures above are placeholders; use your own real prices.) Even businesses that quote every job individually can usually publish a typical range and the factors that move it. That alone turns "pricing unavailable" into an accurate, useful answer.
Step-by-step: publish pricing AI can read
- Find out what AI says now. Ask ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode "How much does [brand] cost?" and "[service] cost in [city]". Record the figures and the sources cited.
- Decide your disclosure level using the table above. If sales objects to fixed prices, agree on ranges or "from" prices.
- Write the key sentence first. Open the pricing page with one plain statement: what it costs, for whom, and on what basis.
- State every condition in text: currency, per user / per month / per project, billing frequency, tax, minimum term, setup fees, regional differences.
- Use real HTML tables for plan comparisons, with text values rather than tick icons. See our guide to AI-friendly tables and lists.
- Explain what changes the price in a short list of factors with typical effects.
- Add structured data where it fits (Product/Offer for products) and keep it identical to the visible prices.
- Date it. Show "Prices last updated: [month year]" and update when prices change.
- Clean up old sources. Update or redirect outdated pricing pages, PDFs and blog posts, and ask directories and review sites to update their listings.
- Re-test monthly and log changes; see our guide to fixing wrong information about your brand in AI.
Structured data for prices
For products sold on your site, Google's merchant listing documentation lists price (or priceSpecification.price) and priceCurrency as required Offer properties, and notes that merchant listings require an Offer rather than an AggregateOffer because the merchant has to be the seller. Product snippets, for pages where people cannot buy directly, accept either. Schema.org's price specification properties include minPrice and maxPrice for ranges and validFrom/validThrough for time-limited prices. Our guide to merchant listings and product schema covers implementation, and ecommerce brands should also review OpenAI product feeds and the Google Shopping Graph.
Whatever you mark up, the visible page comes first. Google's AI features documentation lists this among its best practices:
“Making sure your structured data matches the visible text on the page”
— Google Search Central, AI features and your website, AI features and your website (Google Search Central)
What a strong pricing page includes
| Element | Why it helps AI answers |
|---|---|
| One-sentence price summary at the top | Gives a quotable, self-contained answer |
| Currency, period, tax status beside each price | Prevents misquotes such as monthly vs annual |
| Plan or package table in HTML | Lets tools match features to prices correctly |
| Price factors list | Explains variation instead of leaving a gap |
| Who each option suits | Matches multi-condition prompts ("for a team of 10") |
| "Last updated" date | Signals freshness; helps you spot stale copies |
| Short pricing FAQs | Answers follow-ups: contracts, refunds, discounts |
For conversion advice on the same page, see pricing page optimisation. Key headline prices also belong on your brand facts page so there is one authoritative reference.
Common mistakes
- "Contact us" with no context. At minimum, publish a range and the factors behind it.
- Prices only in images, PDFs or calculators. Put the numbers in HTML text too.
- Missing conditions. "£49" without "per user per month, billed annually, excl. VAT" will be misread.
- Old prices left live. Outdated pricing pages, blog posts and directory listings keep getting cited.
- Schema that disagrees with the page. Keep markup and visible prices identical.
- Bait "from" prices. A "from" price nobody can actually buy erodes trust when the real quote arrives.
Related Guides
- Customer Reviews and AI Recommendations: How Review Data Shapes Answers
- Podcasts and AI Search: Making Audio Content Citable
- Case Studies AI Can Cite: Structure, Numbers and Proof
Frequently Asked Questions
Why does ChatGPT say our pricing isn't available?
Usually because your site does not state prices or ranges in a form it can find. Publishing clear prices, ranges or typical examples in HTML text gives it accurate material to cite.
We quote every job individually. What can we publish?
A typical range, a 'from' price for the most common job, worked examples of recent projects and the factors that change the price. Each is more useful than 'contact us' alone.
Does pricing schema help AI tools?
Product and Offer markup helps search engines understand product prices, and Google's AI features use Search. Make sure markup matches the visible price; other AI tools do not document using it.
How do I fix a wrong price in AI answers?
Find the cited source, update or remove outdated pages on your site, ask third-party sites to correct listings, and make your current pricing page the clearest source available.
Will publishing prices hurt lead quality?
Many businesses find it filters out poor-fit leads. Ranges and examples let you share expectations without committing to a fixed figure.
Give AI the right numbers
If you do not publish pricing information, AI assistants will either say nothing useful or repeat someone else's figures. Publish what you can, whether prices, ranges or examples, with clear conditions, and keep it current. If you want help restructuring pricing and product pages for search, AI answers and conversions, our conversion optimisation and SEO teams can help. Request a custom quote to talk it through.
References
- Semrush: What is query fan-out? (17 August 2026)
- ChatGPT search (OpenAI Help Center)
- Merchant listing structured data (Google Search Central)
- Intro to Product structured data (Google Search Central)
- schema.org: UnitPriceSpecification (inherits PriceSpecification properties)
- AI features and your website (Google Search Central)
