An image is worth a thousand words – but extracting those words can be a royal pain! How many times do clients have excellent product images but lackluster descriptions, or even missing color/material/pattern data? We're pleased to announce that VersaFeed clients can now use AI/ML tools (including ChatGPT Vision technologies) to extract information directly from their product images.
Here's an example where we ask AI to describe the color and pattern of the product image alone (not passing the AI any text details):
- In the first example, the AI returns "Red with gold floral brocade pattern": an excellent description of both the color (Red/Gold) and the pattern "Floral Brocade". It would be trivial to map this data into a feed for "color" and "pattern" attributes, thereby helping channels like Google/Meta/TikTok surface the product under relevant search terms.
- The second example is equally impressive: the AI notes "Red and black with circular gradient dot pattern". It's easy to see how this technology could really help clients who have good product images but may have substandard text/attribute descriptions in their feeds.
As a final example (not pictured above), we asked AI to generate a full product description for the first product (the women's red heel), again only based on the image and nothing else. It created the following:
"Step into elegance with our Red Brocade High Heel Pumps, a perfect blend of classic style and modern flair. These heels feature a luxurious brocade fabric with a vibrant red and gold pattern, guaranteed to add a touch of sophistication to any outfit. Crafted for comfort and durability, the sleek stiletto heel and pointed toe design ensure you'll make a statement at any event."
Considering the AI is only working off a single image, and no text data, the description is very high quality. Image AI tools can unlock a world of new options for your data feeds. Interested in putting this technology to work for your company? Contact VersaFeed today!