- Partial-knowledge search. Enhance AI-driven ecommerce platforms by incorporating partial-knowledge searches for better customer experiences.
- Ephemeral assets recognition. Integrate examination and recognition of ephemeral product attributes like book colors, packaging imagery and lyric website links.
- Connecting catalogs. Improve search results by connecting cataloging applications that account for unique product attributes with their respective ecommerce platforms.
We were driving home from one of our regular trips to the local taco bar when my wife randomly said, “I need to order some vinyl.” Now I’m not one to turn down any opportunity to add to our steadily growing record collection, but it’s rare that Gill mentions buying any. So, I had to ask what prompted the comment.
“I had a great sing along to myself in the car driving home from the office, and then realized we didn’t have those albums on vinyl.”
I asked her which albums she was considering. The list included a couple of Thin Lizzy albums and an AC/DC option — all of which we have on CD, but Gill was correct in that we don’t have them on vinyl. She then concluded the list with “and a Black Sabbath one … but I don’t know which one it is. I think it’s the first with Ronnie Dio, the guy who replaced Ozzy.”
No problem! A quick web search would identify the missing album quickly. But knowing I was going to write this article, I decided to use our unknown album question as a test.
Related Article: Is Generative AI the Future of Product Content Generation?
AI in Ecommerce: From Exclusion Searches to Accurate Results
Rather than use the ubiquitous search engine, I typed the following into our favorite ecommerce platform: “Black Sabbath first album without Ozzy.”
Most ecommerce platforms index their products and serve up the search results based on information about what the product contains. Here I was asking it to find something by exclusion, and not only by exclusion, but by date order of exclusion.
Topping the list of Black Sabbath albums that it served up was 1980’s "Heaven and Hell" — which was the correct answer! (I double-checked using the aforementioned search engine and a couple of music catalog websites). Color me impressed — impressed enough to immediately click the “Add to My Cart” button.
Normally conversations around the use of AI in ecommerce pertains to using it to build a picture of the consumer — including their buying habits and preferred products — and streamlining the purchasing process. All of that data is valuable and useful, and lies in an area where machine learning can help bring efficiencies and improve the customer experience. But that approach is also a very business-centric point of view.
Related Article: Can Conversational AI Improve the Online Retail Experience?
The Partial-Knowledge Search
What about the customer’s perspective? Most ecommerce platforms still function on a taxonomy-based model where items are tagged with data from product information management systems. This relies on the consumer knowing at least the basic facts about the product to find it. But having worked in retail book sales at one point in my career, I can attest to the fact that consumers often know very little about the product they need. “I can’t remember the title, but the book cover was blue, and it had the Eiffel Tower on it.”
I believe this is an area where AI/ML can really offer another level of capability, driving not just natural language search, but what I term a “partial-knowledge search.”
We should strive to incorporate the ability to examine and recognize ephemeral product assets into ecommerce platforms. Those attributes could include book colors, packaging imaging, links to lyrics websites and more.
Several cataloging type applications already do some of this (for instance Goodreads shows that it has 177 books cataloged that have the Eiffel Tower on the cover), but they aren’t usually connected to related ecommerce platforms.
By taking a more customer-centric view of the way that people actually search for products and apply technology to build knowledge not just about the way we shop, but also about the more esoteric ways we look for what we need, we can improve the customer experience even more.
Learn how you can join our contributor community.