Photo by Brando Makes Branding on Unsplash
Over 100 billion visual searches already happened in 2024 through Google Lens and related tools. Yet most websites remain invisible to these searches because search engines and AI systems can't see images the way humans do. They have to read about them.
Alt text and image schema markup are the bridge between what your images show and what Google Lens, ChatGPT, Claude, and Perplexity can understand. These tools control whether your product photos, tutorials, and data visualizations stay buried or unlock visual search traffic, AI citations, and rich results that drive real clicks.
This guide explains how to write alt text that works for both accessibility and search, implement ImageObject schema markup correctly, and structure your images so AI systems cite them. The payoff is measured in traffic, conversions, and visibility where your audience is actually looking.
Key takeaways
- Alt text is the single highest-leverage signal you control for ranking in Google Images.
- Pages that combine text, images, video, and structured data see 156% higher selection rates, with full multimodal plus schema integration delivering up to 317% more citations in AI Overviews.
- Sites using visual search see an 11% average increase in conversion rate.
- Shoppers complete search tasks 38% faster with image input than with keywords.
Why alt text matters more now than it ever has
For years, alt text was treated as a checkbox for accessibility. Now it serves a second reader far more influential: AI systems that decide whether to cite your page.
Alt text is the primary signal Google uses to understand, index, and rank your images in Google Images. ChatGPT, Claude, and Perplexity read your HTML, including the alt attribute, and use it to decide whether your page is relevant enough to cite. When someone asks Claude for product recommendations or ChatGPT for tutorials, the AI doesn't run computer vision on every image. It reads your alt text.
Google Images drives 22.6% of all web search traffic and appears on 37.8% of search result pages, and Google Lens processes 20+ billion visual searches per month. Product images with descriptive alt text get cited more often than images with empty or generic alt text. This isn't just a ranking signal. In AI search, citations drive visibility whether users click through or not.
Retailers implementing visual search see average conversion rate increases of 30%, and the barrier to entry is low. Most competing sites haven't optimized their alt text at all.
Writing alt text that AI systems actually read
The goal is simple: describe what's in the image in plain, specific language a machine can quote.
The formula: be specific, stay concise, provide context
Alt text works for AI search when it describes the image in a clear, specific sentence a machine can quote, not a keyword string.
Describe what's actually in the image using clear, specific language. Think about what someone would need to know if they couldn't see the image. Instead of saying "dog," say "golden retriever puppy playing with a tennis ball in the grass." The more accurate the description, the more useful it is for both users and search engines.
Recommended alt text length is 5–15 words, or under 125 characters. Keep alt text concise enough that screen readers and engines don't lose the thread.
Three rules keep alt text useful:
Describe what is in the frame, not what you wish ranked. Skip "image of" and "photo of," since the engine already knows it is an image. Leave decorative images, spacers, and background textures with empty alt so the engine spends its attention on the pictures that carry meaning.
Real example
Not this:
alt="best blue running shoes cheap"
This:
alt="Side view of Men's TrailRunner Pro in cobalt blue with reinforced rubber tread"
A filename such as brand-productname-blue-leather.webp is more useful than IMG_4091.jpg.
Pairing alt text with schema markup for structured visibility
Alt text tells AI what an image shows. Schema markup tells search engines what role it plays on your page. Together, they unlock rich results and AI citations.
Product schema and Article schema use image descriptions as signals when deciding which images to feature in rich results. But only if the alt text and schema align.
For ecommerce specifically, make sure the image field in your Product schema points to the same image whose alt text is well-written. Mismatches between schema metadata and actual alt text can reduce rich result eligibility.
Multimodal pages with text, images, video, and structured data get selected into AI Overviews at 34.6% rate, a 317% lift over text alone. This shows that a page an AI engine can parse across multiple formats gets picked more often than a page it has to read as plain text.
The fix is straightforward: implement ImageObject schema markup for your images and make sure the caption and description match the alt text in spirit. Don't duplicate word-for-word (that triggers spam filters), but describe the same thing.
Understanding ImageObject schema and when to use it
ImageObject schema is the primary structured data type for image SEO, providing explicit context about image content, licensing, and attribution. By adding structured data to your HTML code, you help search engines and AI systems understand your images with greater accuracy and context.
For images, the primary schema type is ImageObject. It provides context about visual content, helping search engines better understand what an image represents. It can also be combined with Product schema to show pricing, availability, or brand details.
When to use ImageObject
- Standalone photographs or galleries: Use ImageObject to markup each image with description, author, and license information.
- Product images: Nest ImageObject inside Product schema to explicitly connect photos to products.
- Editorial images: Pair Article schema with ImageObject to signal which images are primary for that piece.
- Instructional images: Include ImageObject in HowTo schema to describe each step's visual.
Core ImageObject properties
The essential properties are:
- url (required): The absolute URL to the image file (must start with https://)
- width, height (required): Dimensions in pixels
- caption: A description of what the image shows
- contentUrl: Direct link to the image
- name: Short title for the image
- author: Creator or photographer
- creditText, copyrightHolder: Attribution and rights information
Google prioritizes schema-defined images for rich results and Knowledge Panels, and use high-resolution images (minimum 1200px wide) for best eligibility in Google Discover and rich results.
A minimal JSON-LD example
{
"@context": "https://schema.org",
"@type": "ImageObject",
"url": "https://example.com/images/product-navy-blazer-front.jpg",
"contentUrl": "https://example.com/images/product-navy-blazer-front.jpg",
"width": 1200,
"height": 900,
"caption": "Navy wool blazer with notch lapels and functional button cuffs",
"name": "Navy Wool Blazer - Front View",
"creator": {
"@type": "Person",
"name": "Studio Photographer"
}
}
Implementing schema for ecommerce and editorial
Different content types need different approaches.
For product pages
The image property inside Product schema is what connects your product photos to rich results, free Shopping listings, and AI Overviews.
Nest your ImageObject inside Product schema:
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Premium Wireless Headphones",
"image": {
"@type": "ImageObject",
"url": "https://example.com/headphones-silver-angle.jpg",
"width": 1200,
"height": 800,
"caption": "Premium Wireless Headphones, side angle view showing silver finish and noise-canceling design"
},
"price": "199.99",
"priceCurrency": "USD",
"availability": "https://schema.org/InStock"
}
Multiple product images require array notation in your JSON-LD, and each image should have its own ImageObject markup. This matters most for fashion and lifestyle products, where multiple angles and styling options help drive conversions.
For blog posts and articles
Use Article schema and declare your primary image:
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "How to Repot a Fiddle Leaf Fig Plant",
"image": {
"@type": "ImageObject",
"url": "https://example.com/blog/fiddle-leaf-repot-step-1.jpg",
"width": 1200,
"height": 800,
"caption": "Fiddle leaf fig in ceramic pot being carefully lifted before repotting"
},
"datePublished": "2026-10-01"
}
Common mistakes to avoid
Your schema markup should match what's actually on the page. If your ImageObject schema claims an image is 1920x1080 but the real image is 800x600, search engines will notice and may ignore your markup entirely.
All URLs must be absolute and accessible. Relative paths or URLs pointing to images behind authentication fail validation. Every URL in your markup must be crawlable by search engines.
How AI assistants read your images in 2026
Understanding how different AI platforms consume images shapes how you optimize.
Generative Engine Optimization (GEO) is the practice of making your content discoverable to AI search assistants like ChatGPT, Claude, Perplexity, and Google AI Overviews. For images specifically, GEO is mostly about alt text, captions, structured data, and surrounding text, the things AI assistants actually read.
When someone asks ChatGPT "show me a credenza that fits under a window," the AI assistant pulls candidate pages largely based on textual signals: alt text, product descriptions, surrounding paragraph copy, and structured data. Pages with rich textual context get cited. Pages with empty alt and generic captions don't.
The hierarchy of signals AI systems read:
- Alt text (highest leverage)
- Image captions and surrounding body text
- Schema markup (ImageObject, Product, Article)
- Image filename
- EXIF data (lowest leverage)
Image captions provide an additional layer of context that both users and AI systems use. Captions are read at a higher rate than body text by human visitors, and AI systems treat them as high-confidence descriptions. Use captions to add context that complements (not duplicates) your alt text.
Measuring impact: What actually moves the needle
Visual search isn't a vanity metric anymore. Google Lens now exceeds 25 billion queries per month, with roughly 20% of those searches showing commercial intent.
Aggregated ecommerce data shows sites with visual search see an 11% increase in conversion rate on average. The same dataset shows search users overall convert at higher rates than browsers who don't use search.
For AI citation specifically, sites cited inside AI Overviews can see higher CTR than uncited results on the same SERP, with cited brands earning roughly 35% more clicks than uncited ones. Being visible in the AI answer, even without a direct link, increases click-through rates on your other search results.
Track these metrics in Google Search Console:
- Image impressions and clicks (Google Images)
- Rich result eligibility (structured data reports)
- Visual search traffic through utm parameters or referrer data
Set up Google Analytics to isolate image-driven traffic. Most shops don't, which means they're flying blind on one of the fastest-growing channels.
Quick implementation checklist
Here's the fastest path to optimization:
- Audit your top 50 product or content pages for empty alt text
- Rewrite alt text to be specific, under 125 characters, and descriptive of what's actually shown
- For product pages, add ImageObject schema nested in Product schema
- For blog posts, add ImageObject inside Article schema
- Ensure all image URLs in schema are absolute and crawlable
- Match schema captions conceptually to alt text (don't duplicate exactly)
- Validate schema using Google's Rich Results Test
- Monitor Google Search Console for image indexation and rich results eligibility
Start with your money pages, the product pages or content pieces that already drive revenue or traffic. The lift on high-value pages justifies the effort. Then work outward to lower-traffic content.
FAQ
Should I stuff keywords into my alt text?
No. Focus on creating useful, information-rich content that uses keywords appropriately and in context. Avoid filling alt attributes with keywords (keyword stuffing) as it results in a negative user experience and may cause your site to be seen as spam. Write naturally. If the keyword fits, great. If it doesn't, leave it out.
Does schema markup actually help images rank?
Indirectly but meaningfully. Schema doesn't rank the image file itself, but it does attach your photos to rich results, Shopping listings, and AI-generated answers. Think of schema as a citation amplifier, not a ranking multiplier.
What's the minimum size for images in search results?
Use high-resolution images with a minimum of 1200 pixels wide for best eligibility in Google Discover and rich results. Smaller images still index but have a lower chance of being selected for rich results or being cited by AI systems.
Can I use the same alt text for multiple images?
Only if the images are genuinely identical. If you have multiple product angles or variations, each image needs its own description. Each image should have its own ImageObject markup. This matters most for fashion and lifestyle products, where multiple angles and styling options help drive conversions.
How often should I update alt text on old content?
While addressing every image would be ideal, it may not be practical. Begin by auditing your site to identify the most visited or SEO-critical pages and prioritize adding or improving alt text there first. Then gradually work through older or lower-traffic content.