Multimodal AI
Multimodal AI is AI that works with more than text, like images, audio and video, so it can read a photo or answer a question about a chart.
Multimodal AI is AI that can take in, and sometimes produce, more than one kind of content: text, images, audio and video. You can show it a photo of a coffee machine and ask which model you're looking at, or give it a chart and ask what it shows.
For brands, it means your images and videos can be understood by the AI that reads your pages, and that customers can start a search with a picture instead of words.
What it changes
Many of the main AI assistants can now take images and voice as input, alongside text. That opens up new ways in:
- A shopper takes a photo of a bag of coffee in a café and asks where to buy it
- Someone uploads a screenshot of a product page and asks if the price is fair
- A person asks out loud, in the car, for a bakery near them that's open now
For example, someone photographs a Kestrel Roasters bag at a friend's house and asks "what's this like, and is there something similar that's cheaper?" The answer depends on the assistant recognising the bag, knowing the product, and finding alternatives.
What it means for your content
Make your products easy to recognise. Use clear photos of your packaging and logo, and keep them consistent. Write descriptive alt text and captions, so the meaning is there in words as well. Add transcripts to your videos.
Don't rely on images alone for facts that matter, though. A price in a banner image may be readable to some models, but text in the page's HTML is the safe bet for every bot that reads it.