For years, image search was separate from text search. Google Image Search was its own thing. You optimised images for image search and text for web search. They were parallel systems. Multimodal AI changes this. Systems like Claude and the latest versions of other AI models can understand and search across text, images, and video simultaneously. A query can include images. Answers can cite images. Your visual content is no longer optional—it's part of your discovery strategy.
This matters for two reasons. First, many discovery queries are now multimodal. Someone might show an AI an image and ask "where can I get something like this?" or "what is this and who makes it?" AI systems can understand the image and find sources. If your products or services appear in images across the web, you're discoverable. Second, people are increasingly using visual search because it's faster than describing what you want in words. If your visual content isn't discoverable and isn't properly attributed, you're losing an entire discovery channel.
How Multimodal Search Works
Multimodal search allows queries that combine different types of input. You can ask about an image you're looking at. You can ask an AI to find images that match text descriptions. You can upload a photo of a space and ask the system to recommend furniture that would work in that space. The system understands both the visual and textual information and reasons across them.
For businesses, this means your visual content—product photos, before-and-afters, portfolio images, team photos—is now part of your discovery surface. Someone can search for something visually similar to your work and find you. Someone can ask for recommendations that match a visual description and your images might show up. This is new. In traditional search, your images were indexed by Google Images, but they weren't part of the main text-based search discovery.
The implications are that image quality matters for discovery now. A blurry, poorly lit photo is less likely to be found and cited than a professional, well-lit image. An image with clear branding is more useful for discovery than an ambiguous image. Images with context matter more than isolated images. If you're posting images of your work online, you're contributing to your discoverability in multimodal search. The better your images, the better your discovery.
Image SEO and Attribution
Images still need basic SEO. File names should be descriptive. Alt text should be accurate and complete. File sizes should be optimised for web. Metadata should include useful information. These were always good practices, but they're more important now because AI systems use this information to understand what images are about. A photo with alt text that says "image" doesn't help. A photo with alt text that says "luxury residential bathroom with marble counters and heated floors" helps the system understand what the image shows and when it's relevant.
Attribution matters more in multimodal search. If an image appears across the web without clear attribution to your business, you won't get credit for it. Make sure your images have watermarks or clear branding. Make sure the pages where your images appear have clear information about your business. If someone uses your image elsewhere, you have a claim to it. If your image appears in contexts where you're clearly the creator or source, you get discovery credit.
Visual Consistency and Branding
As multimodal search grows, visual consistency becomes part of your discovery strategy. If you have a consistent visual style—consistent colours, consistent fonts, consistent photography style—systems learn to recognise your work. When someone searches for something visually similar, your consistent style makes you more findable. If your visual style is all over the place, you're harder to recognise.
This doesn't mean all your images need to look identical. It means developing a visual brand that's recognisable. Maybe it's a consistent colour palette. Maybe it's a consistent photography style. Maybe it's a consistent design approach. Whatever it is, consistency helps multimodal systems understand that a collection of images is from the same source. It helps you get discovery credit across multiple images.
Video and Multimodal Search
Video is becoming searchable in multimodal systems. AI systems can understand and search video content. A person can search for "how to install ceiling fans" and find relevant video content. A system can recommend a video tour of a property. Video creators who optimise their content with good descriptions, clear visuals, and proper attribution will be discoverable.
For businesses, this means video content is increasingly important. A tour of your facility. A demonstration of your service. A walkthrough of your process. These videos become discoverable when properly optimised. Transcripts help. Proper titles and descriptions help. Good lighting and clear visuals help. If your video is high quality and properly described, it becomes discoverable in multimodal search.
Building a Multimodal Content Strategy
Your content strategy now needs to include visual content as a primary element, not a secondary one. High-quality photography of your products or services. Professional videography showing your work. Infographics that communicate your expertise. All of these contribute to your multimodal discoverability. You're not just writing for text-based discovery. You're creating visual assets that can be found and cited independently of text.
This means investing in photography and videography if you haven't already. Professional photos are more discoverable than phone photos. Professional video is more discoverable than amateur video. This isn't just about aesthetics. It's about discoverability. If your images and videos are professional, well-lit, and clearly represent your work, they're more likely to be found, cited, and attributed in multimodal search.
It also means being strategic about where your visual content appears. Appearing on your own website is good. Appearing in publications and media is better because it increases the distribution of your visual content and the likelihood it will be cited. If your work is featured in design magazines, blog posts, or industry publications with full credit to you, multimodal search will find those images and connect them back to your brand.
The Integration Point
Multimodal search is where text discovery and visual discovery converge. Your website text and your visual content are now part of the same discovery system. They reinforce each other. Great photos with proper attribution and context support your text discovery. Great text that's illustrated with high-quality images supports your visual discovery. The two work together. If you're missing one, you're only getting half the benefit.
This is the last major shift in search strategy. By 2028 or 2029, multimodal search will be standard. Systems will expect text and images together. Businesses that are optimising for text-only discovery will be leaving significant visibility on the table. Businesses that have invested in both text content and visual content will dominate. The time to start is now.
— Sam