Our Data Sucks! GenAI Won’t Work for Us.
The e-commerce search bar is about to undergo the biggest transformation since platforms like Shopify revolutionized online shopping. As generative AI (GenAI) conversational shopping capabilities are being integrated into search, the way we shop online will never be the same.
October 8, 2024
Epiphany moment for me this week. Sometimes when you’re deep in the woods, you forget that not everyone is in the woods hanging around with you, singing Kumbaya.
Limited Data
I believe that a material number of retailers do not feel that their product catalog has enough data associated with each product to take advantage of adding GenAI natural language or precise color search, like that offered by shopperGPT.
Real estate for online e-commerce is not limitless. You have seconds to make your point with a shopper. So most products consist of a product name, a short bullet list of features and maybe 4 images. Many sites also include a text-based color description or palette of the primary colors. How could that be used to respond effectively to a natural language query like this:
“I’m searching for a maxi skirt that’s perfect for a casual day out in the Fall here in San Francisco. It should be made of a lightweight material like cotton or linen, and in a solid color or a subtle print. I’d prefer it to have a high waist and a flowy silhouette. Oh yeah, and I’d like to keep it under $50, please.”
Wait…what? How the heck to I find that with the above product data? Not possible, right?
GenAI to GenAI
Not exactly. Trade secret reveal alert – we use GenAI to set up sites for GenAI! On data inspection and ingestion, we enhance and augment a site’s data with everything we can glean form the images and existing content.
How we do that and what we do after that is AI-complex (and proprietary!), but suffice it to say that when the data ingestion process is complete, we can deliver back the closest possible matches in the catalog to the above query 10-20x faster than if a shopper tried to navigate a site starting with a “Maxi Skirt” text query. Same is true for finding the closest matches to an exact shade of green (mint perhaps?).
And by the way, that’s generally in 6 seconds or less vs. 3 minutes or more. (LinkedIn, we need emojis for statements like these ; )
So as I’ve stated in the past, GenAI is coming to every popular e-commerce site’s search bar. And with the KPI upside that massively accelerated product discovery promises, I’d say much sooner than later.
visualAI retail solutions specializes in leveraging advanced AI technologies to enhance the retail shopping experience, providing tools and APIs like shopperGPT for AI-powered natural language search, color-based search, similar image search, and hyper-personalization, enabling seamless and intuitive product discovery for customers.