AI Powered Search
Increase Accuracy of Search Mechanism on eCommerce Platform to Increase Conversion Rate and Drive Increased Sales.
Traditionally search bars use keyword frequency to find objects of similarity. We built a solution that is fundamentally different and dramatically more accurate using Google Gemini’s Enterprise Product Search solutions. The Artificial Intelligence solution utilized embeddings which transformed our client’s ecommerce site search query and their product catalog into a multi-dimensional matrix. This means that we transformed all metadata into numerical transcriptions that describe a physical location. When a user searches the ecommerce website, we search the physical matrices for products close in proximity and respond back to the site with similar products that are much more relevant to the customer.
Our AI Powered Search Feature was intentionally not tested due to the dramatically better performance and instead was immediately rolled out to production. While we do always advocate for A/B testing, some dated solutions are quickly being shut off regardless. Based on our research of similar solutions that had been A/B tested elsewhere, we found that AI search can provide up to a 216% increase in conversion rate compared to traditional search.
