"Image search for dinnerware is something we’ve been investigating for years. The searches are complicated by the non-distinctive shapes and many collections with similar colors and designs. We’ve attempted image search with several groups over the years, but it wasn’t until we initiated a pilot project with Grid Dynamics that we were able to see results. Their experience with Machine Learning and the Google Cloud tools allowed them to successfully build a ML model that worked for matching dinnerware patterns. Even after our initial success with dinnerware, I was shocked when they were able to demonstrate a model that worked for crystal. For the most part, crystal has similar shapes, clear glass color, and intricate details that are hard to distinguish in images. Over the years, they have advanced our models from dinnerware and drinkware to other product lines like silver-flatware and collectibles. The collaborative product provides any person the ability to identify their existing items without needing to be an expert.”