Oliver Tan is founder/CEO of ViSenze
By Oliver Tan After years of R&D in artificial intelligence and machine learning by technology innovators, the first deployments of visual search are going live. And as applications of machine learning are growing more mature, fashion retailers are among the first to profit. Why now? AI on the prize During recent years, we have seen three important technology trends coalesce to make the promise of artificial intelligence a reality. First, software is widely available that can gather and organize massive amounts of data. Next, even inexpensive computers now have impressive processing power. Finally, the price of storage has dropped dramatically. Combined, these developments have enabled startups to create powerful solutions using artificial intelligence. It seems that every week we hear of another fashion or accessory retailer closing stores or declaring bankruptcy. The bricks-and-mortar business model does not offer the financial opportunity it once did. Shopping for clothing, handbags, shoes and jewelry has moved online faster than retailers have been able to adapt to the change. They tried countless ways to replicate their offline strategies or – in the worst of cases – deferred to search engines to make their products available over the Web. The results have been disappointing. I believe that retailers have not yet been able to extract the full financial potential of online selling because they are depending on text-based or keyword search for a visually driven industry. In 2017, retailers finally have a tool to help them retake control of their destiny. That tool is visual search. AI still has much to learn That was the good news. The bad news is that artificial intelligence (AI) still has much to learn and learning requires data, data and more data. Earlier this year, Pinterest launched “Pinterest Lens.” The app, which is still in beta, considers the shape, color, product category and more to make relevant suggestions to users. Pinterest is the first to say, however, that its Lens is “not perfect.” Pinterest Lens is one of the most high-profile launches of visual search to date, but not the first. Amazon’s app allows users to upload a picture of an object and then get a price from its “Everything Store.” Shutterstock and Google offer “reverse image search tools.” Instagram and Facebook are also quickly innovating to improve their understanding of the content within images. Clearly, the interest in visual search is huge and growing fast. However, these artificial intelligence solutions do not cater to retailers – the category most primed to benefit. AI solutions must focus The problem as I see it is that Pinterest, Amazon and the others are trying to do too much. I believe the fastest way to make artificial intelligence smarter and visual search profitable is to focus learning in specific areas. If you have spent any time on Pinterest, you know that you can find images and information about absolutely anything, from cute animals to formulas for black holes. This is Pinterest’s strength, but when it comes to visual search, it is also its weakness. It will take time and billions of images for Pinterest Lens to consistently deliver compelling results – i.e. “Shoppable Pins” – and revenue for retailers. This massive need for data is undoubtedly the reason why Pinterest recently signed a deal with Samsung to be the visual search technology behind Samsung’s Bixby virtual assistant. Access to the data from Samsung users – Samsung phones now constitute 26.1 percent of the market – could be a windfall for Pinterest. In the end this deal may prove better for Pinterest to train their machine learning systems than for Samsung’s customers. ALL THAT SAID, AI and visual search offer a major financial opportunity for retailers in the fashion and accessories industry. Yet, while it is early days for artificial intelligence, there is no time for retailers to waste. This year offers an opportunity to gain an edge on the competition by providing today’s Web-savvy and highly visual shoppers with a better, faster way to find what they want. Oliver Tan is founder/CEO of ViSenze, San Francisco. Reach him at oliver@visenze.com.