A product image is optimized for AI shopping when it shows the exact product, clearly, in a file a system can fetch, with a name and a description a system can read, and with the same facts in the structured data that the shopper sees on the page. A good image does not sell the product on its own. It lets a system verify what the product looks like, and lets a shopper trust what the system says about it.
Google says that high quality images are part of valuable, non-commodity content, and that if you are already following its image SEO best practices, you are already optimizing for generative AI search. It also says that no special markup is required for generative AI search. The work is the image itself, the words around it, and the markup that states what it shows.
Where it goes wrong
Brands use images that look good in a catalog and fail on a product page, because the image does not show the product a shopper will receive.
- The main image is a lifestyle shot. A lifestyle shot shows a mood, not a product. A shopper who clicks through expecting the product in the picture, and finds a different color or a different size, will leave, and a system that reads the page will learn not to trust the next image.
- The image is too small, or it is a thumbnail sent as the main image. Google Merchant Center says to provide images at least 500 by 500 pixels, and recommends around 1500 by 1500 pixels for best performance. A small image cannot be checked, and a thumbnail sent as the main image is a sign the store did not check its own feed.
- The image has promotional text, a watermark, or a border. Google Merchant Center says not to include promotional text, watermarks, or borders, and not to submit a placeholder or a generic image. A product image that covers the product with a sale banner is not a product image.
- The image has a generic filename and no alt text. Google says to use filenames that are short but descriptive, and that alt text is the most important attribute for providing more metadata for an image. A file named image1.jpg with no alt text tells a system nothing about what it shows.
- The image in the structured data does not match the image on the page. A Product block that points to one image, while the page shows another, teaches a system not to trust the block, and the next fact in the block will be doubted too.
The steps
Step 1. Show the exact product, whole, on a clean background. The marketer picks the main image that shows the entire product, without staging, without a prop that hides it, and without a promotional overlay. The image must clearly display the exact item being sold. If the product comes in colors, the main image must show the color the page is selling, not a different color.
Step 2. Meet the size and format rules. The developer checks the image against the Merchant Center rules: at least 500 by 500 pixels, around 1500 by 1500 pixels recommended, no image larger than 64 megapixels, no file larger than 16 MB, and a supported format, JPEG, WebP, PNG, GIF, BMP, or TIFF, with the file extension matching the format. Do not scale up a small image, and do not submit a thumbnail as the main image.
Step 3. Give the image a descriptive filename. The marketer renames the file so it says what the image shows, in short plain words, such as trail-running-shoe-blue.jpg, not IMG00023.jpg. Google says a filename can give very light clues about the subject matter of the image, and that generic filenames should be avoided where possible.
Step 4. Write alt text that describes what the image shows. The marketer writes alt text that says what is in the image, in useful, information rich words, in context with the page. Google says it uses alt text along with computer vision and the contents of the page to understand the subject matter of the image, and that keyword stuffing in alt text is spam. Write for a person who cannot see the image, not for a search engine.
Step 5. Put the image near relevant text, on the right page. The marketer checks that the image sits on the product page for that product, near the text that describes it. Google says to place images near relevant text and on pages relevant to the image subject matter. An image on a category page, or on a page for a different product, will be read as a different product.
Step 6. State the same image in the structured data and in the feed. The developer sets the Product image field to the same image URL the page shows, and sets the Merchant Center imagelink attribute to the same URL. Google Merchant Center says imagelink is required for each product, and that the URL must be crawlable by Googlebot and Googlebot-image. Check that robots.txt allows both, and that the URL starts with http or https.
The machine-readable piece
This is the shape of a Product block with an image, using the schema.org vocabulary. Replace the placeholders with your own product facts, and only use values that match the visible page. The image URL must be the same image the shopper sees on the page.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Example Wooden Dining Chair",
"image": "https://[yourdomain.com]/images/wooden-dining-chair-oak.jpg",
"description": "A wooden dining chair in oak, with a woven seat and a straight back, sold as a single chair.",
"brand": {
"@type": "Brand",
"name": "Example Furniture Brand"
},
"offers": {
"@type": "Offer",
"price": "249.00",
"priceCurrency": "USD",
"availability": "https://schema.org/InStock"
}
}
</script>
This is the shape of the same image in a Merchant Center feed row. Replace the placeholders with your own values, and keep the image URL identical to the one in the block above.
id, title, link, image_link, availability, price
SKU-CHAIR-OAK, Example Wooden Dining Chair, https://[yourdomain.com]/products/wooden-dining-chair-oak, https://[yourdomain.com]/images/wooden-dining-chair-oak.jpg, in_stock, 249.00 USD
Check both after you publish them. Open the page and confirm the image matches the block and the feed. Then run the page through the Rich Results Test and fix any error it reports. An image that matches in all three places, page, block, and feed, is an image a system can trust.
What it looks like when it works
Say you run a furniture brand, selling dining chairs, tables, and sideboards, with a catalog of about sixty products. The marketer checks the main image on every product page. The dining chair page shows a lifestyle shot of a full dining room, with the chair half hidden behind a table. The marketer replaces it with a clean shot of the chair alone, in oak, on a plain background, at 1500 by 1500 pixels, named wooden-dining-chair-oak.jpg, with alt text that says a wooden dining chair in oak with a woven seat.
The developer sets the Product image field and the Merchant Center image_link to the same URL, and checks that robots.txt allows Googlebot-image. A shopper asks an assistant for a wooden dining chair in oak, and the assistant can find the page, read the image and the text together, and cite the brand with the product as shown. The image did not make the sale on its own. It showed the exact product, clearly, in a form a system could verify.
The fundamentals
These mechanics are shared across every engine and covered in full in their own guides:
- •The llms.txt format and a complete sample, see our guide to writing a good llms.txt file
- •The markup AI systems read to verify facts, see our product schema guide
- •The complete crawler rule reference, see our robots.txt guide for AI search
What this guide does not cover
This guide covers how to optimize product images for AI shopping: the image itself, the filename, the alt text, and the structured data and feed. It does not cover how to shoot product photography, the full technical audit that checks every crawler and every page, or the transaction layer that lets an agent complete a purchase. It does not cover how to run a Merchant Center account in full. Those are separate jobs, and this guide is 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, Optimizing your website for generative AI features on Google Search (specific): https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
- Google Search Central, Image SEO best practices (specific): https://developers.google.com/search/docs/appearance/google-images
- Google Search Central, Product structured data (specific): https://developers.google.com/search/docs/appearance/structured-data/product
- Google Merchant Center Help, Image link [image_link] (specific): https://support.google.com/merchants/answer/6324350
- Schema.org, Product type (specific): https://schema.org/Product
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.
