A sale price is only useful if the price an AI system quotes is the price a shopper pays at checkout. That happens when the sale price, the regular price, and the sale window agree in every place a system can read them: the product page, the structured data on that page, and the product feed. If one of those places still shows the old price, or shows the sale price with no end, the assistant can quote a price your store no longer honors, or keep quoting a sale price after the sale has ended.
The fix is a routine, not a trick. Set the sale price in one record, state the regular price beside it, give the sale a start and an end, and update all three places on the same day. No vendor publishes a fixed time for an AI model to pick up a price change, and this guide claims none. What you control is whether the new price is the same everywhere a model can look.
Where it goes wrong
Stale sale prices in AI answers almost always trace to one of these failures, not to a model choosing an old price on purpose.
- The page had changed, but the markup had not. The visible price says one thing, while the structured data block still carries the old price. An AI system that reads the markup will quote the price it finds there, even when it is old.
- The sale price went into the price field, and the regular price vanished. The feed price must stay the regular price, with the sale price submitted beside it. When the regular price is replaced by the sale price, there is no regular price left to compare against, and the sale cannot be shown as a sale.
- The sale has no end. A sale price with no effective date keeps applying until someone removes it by hand. A sale that was meant to run for a few days keeps being quoted for weeks.
- The sale price sits in a banner or a guide, but not in the markup or the feed. A banner can disappear while the page still shows the old price, and a guide that names a price becomes wrong the day the price changes.
- The sale price is lower than the regular price by too little, or the regular price was never charged. Google documents that a sale price must be lower than the regular price, and that the regular price must have been valid for a set period before the sale. A discount that does not meet those rules may not show as a sale.
- The sale had ended, but the sale price was still there. When a discount has ended on the page yet still appears in the markup or the feed, it keeps being quoted until it is taken out of every place.
The steps
Step 1. Set the sale in one record. The marketer sets the regular price, the sale price, the sale start, and the sale end in the product record, once. Every other copy, the page, the structured data, and the feed, should read from that record, or be regenerated from it. If any copy is typed by hand, separately from the record, it will drift. Drift is how an AI answer quotes a price you no longer honor.
Step 2. Show both prices on the page, in plain words. The marketer edits the product page to show the sale price a shopper pays, the regular price crossed out beside it, and the sale end in plain words. The price must be visible on the page, not only after a shopper picks a size or an option, and not only inside an image. An AI system reads text. A price inside an image, or behind a click, is a price it may never see. Check each variant, because a sale price often applies to some variants and not to others.
Step 3. Put the sale price in the structured data, with its end. The developer updates the Offer block so the active price is the sale price, the regular price is marked as the strikethrough price, and the sale end is stated. Google documents that the active price is the amount currently offered, that a strikethrough price is the higher regular price shown during a sale, and that the sale duration should state a start and an end, with a time and a time zone where possible. Use either validThrough or priceValidUntil for the end, never both on the same node. A listing may not display if priceValidUntil points to a past date, so remove or update it when the sale ends.
Step 4. Put the sale in the feed the same day. The marketer or the developer updates the feed row for every product on sale: the price field keeps the regular price, the sale price field carries the sale price, and the sale price effective date carries the start and end with a time and a time zone. Google documents that the sale price must be submitted in addition to the price field set to the non-sale price, and that it must match the landing page and the checkout pages. If you do not submit an effective date, the sale price always applies. If your feed updates on a schedule, plan the sale launch around the feed schedule you actually run.
Step 5. Sweep the rest of the site for the old price. The marketer searches the store for the old price and the old sale wording: guides, comparison pages, collection descriptions, blog posts, and the shipping and returns page where it mentions a price. Fix each one, or take the price out of it. A guide that prints a price is out of date the moment the price moves, so have guides link to the product page for the current price instead of printing it.
Step 6. Close the sale the same way you opened it. When the sale ends, the marketer takes the sale price off the page, the developer takes it out of the markup, and the feed is updated that same day. Once the sale has ended, open the page and check that only the regular price is showing. The usual reason a store keeps quoting a price it no longer charges is a sale price that was never taken down after the sale ended.
The machine-readable piece
This is the Offer block for a product on sale. Swap the placeholders for your own values. Those fields come from the schema.org Offer type and the Google Product structured data page. The active price is the sale price, the regular price is marked as the strikethrough price, and the sale end is stated with a time and a time zone. Price, currency, and availability must match the visible page exactly.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Example cycling jacket",
"offers": {
"@type": "Offer",
"price": "89.00",
"priceCurrency": "USD",
"availability": "https://schema.org/InStock",
"validFrom": "YYYY-MM-DDTHH:MM:SS+HH:MM",
"priceValidUntil": "YYYY-MM-DDTHH:MM:SS+HH:MM",
"priceSpecification": {
"@type": "UnitPriceSpecification",
"price": "119.00",
"priceCurrency": "USD",
"priceType": "https://schema.org/StrikethroughPrice"
},
"url": "https://[yourdomain.com]/products/example-cycling-jacket"
}
}
</script>
For the Merchant Center feed, the matching row carries the regular price, the sale price, and the sale window. The fields below come from the Google product data specification. That row is tab-delimited, with tabs between the values. Swap every placeholder for your own values.
id title link price sale_price sale_price_effective_date availability
example-jacket-01 Example cycling jacket https://[yourdomain.com]/products/example-cycling-jacket 119.00 USD 89.00 USD YYYY-MM-DDTHH:MM+HHMM/YYYY-MM-DDTHH:MM+HHMM in_stock
Check both after every sale change. View the page source and confirm the block carries the sale price and the sale end, then open the feed and confirm the row carries the same prices and the same window. Where the two disagree, fix the one that is wrong. Never add a field you cannot fill with a true value, and never guess a value just to fill a row.
What it looks like when it works
On the page, a shopper sees the sale price next to the size options, the regular price crossed out beside it, the sale end in plain words, and the stock status in words. Under the surface, the structured data block carries the same sale price and the same end, and the feed row carries the regular price, the sale price, and the same window. To a human reader, nothing on the page looks different. The change is that every place a machine can look states the same numbers.
Say you run a cycling apparel store, selling jackets, jerseys, and shorts, with a sale on jackets for a few days. The marketer sets the regular price, the sale price, and the sale window in the product record, edits the page to show both prices and the sale end, and tells the developer the same day. The developer brings the structured data block into line and checks the page source again. The marketer updates the feed row for each jacket, then searches the site and finds the old price still printed in a comparison guide, and changes the guide to link to the product page for the current price instead of printing it. A shopper asks an assistant for a cycling jacket under a set budget. The assistant reads the page, the markup, and the feed, finds the same sale price in all three, and quotes it. When the sale ends, the marketer puts the regular price back on the page, the developer puts it back in the markup, and the feed is updated that same day, so the sale price stops being quoted.
The fundamentals
These mechanics are shared across every engine, and their own guides cover them in full:
- •The llms.txt format, with a full sample, is in our guide to writing a good llms.txt file
- •The markup AI systems read to verify facts is covered in our product schema guide
- •The full crawler rule reference lives in our robots.txt guide for AI search
What this guide does not cover
This guide covers keeping sale prices current in the places AI systems read. It leaves out the full technical audit across every crawler and every page, the transaction layer that lets an agent finish a purchase, and the per-engine playbooks for each assistant. It does not cover pricing strategy or discount design. Those are separate jobs, and this guide is only the map, not the whole trip.
Check your own store now
Run the free scan. One click, no signup, and it shows whether AI crawlers can read your store today: aizii.ai/aeo-scan
Run the free scanSources
- Google Search Central, Product structured data (specific): https://developers.google.com/search/docs/appearance/structured-data/product
- Google Merchant Center, Product data specification (specific): https://support.google.com/merchants/answer/7052112
- Google Merchant Center, Sale price effective date (specific): https://support.google.com/merchants/answer/6324460
- Google Search Central, Optimizing your website for generative AI features on Google Search (specific): https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
- Schema.org, Product type (specific): https://schema.org/Product
- Schema.org, Offer type (specific): https://schema.org/Offer
Note: links marked specific point to the page that supports the claim in this guide. Links marked general docs point to the vendor's general docs section for that topic, not to a page that proves every step. Check the vendor docs for the current path before relying on a menu name or click path in this guide.
