This interview is part of our new AI in Retail series, where we interview the world's top thought leaders on the front lines of the intersections between AI and retail.
In this interview, we speak with Anastasia Sartan, Founder and CEO of StyleHacks, to understand how her company is using AI to transform retail, and what the future of retail holds.
1. What’s the story behind StyleHacks? Why and how did you begin?
AS: Before StyleHacks, I founded the first and one of the largest fashion e-tailers in Russia and realized that the way the fashion industry worked was not sustainable. The industry has been centered around brands and trends, forcing the customers to speak its language instead of speaking theirs. So, three years ago I founded StyleHacks to empower customers to take back control of their style and wardrobe, and to empower brands to sell better, more effectively, and more responsibly. With StyleHacks, you can just say what you are looking for like you would to a friendly store associate, and find the perfect piece in minutes, absolutely hands-free. We received funding from the Google Assistant Investment Program and Founders Fund and were prominently featured at 2017 F8 and Vanity Fair’s Founders Fair.
2. Please describe your use case and how StyleHacks uses artificial intelligence.
AS: Last year, Google reported a 120 percent increase in ultra-personalized searches. Instead of “sneakers,” people started searching for “running shoes for overpronation.” At the same time, we are observing the rise of voice-assisted devices: today, 26.2 percent of all U.S. adults have access to a smart speaker, and it is estimated that in ten years most products will be purchased via voice-activated interfaces.
We have put the two together before anyone else, allowing retailers to satisfy these requests with our technology. It consists of three modules that employ an array of AI tools to deliver ultra-personalized results to the customers instantly: NLU (Natural Language Understanding) module, Expert Rules module, and Inventory Navigation module.
- NLU uses Google DialogFlow and custom code to understand what the user said and translate it into a specific structured request.
- Expert Rules implement expert knowledge translating “fuzzy” requests into the structured filters from the given domain. (e.g. “jeans for work” translates to “not distressed, dark color, etc.“).
- The Inventory Navigation module is a software component that employs DeViSE (Deep Visual-Semantic Embedding) ML model for a semantic search along textual and visual representations of the product in the multidimensional vector space. Our models are trained on the custom corpus of fashion texts.
StyleHacks is a platform-agnostic product that could be adapted to different conversational mediums, like Voice Assistants, Chatbots and such.
3. Could you share a specific customer/user that benefits from what you offer? What has your service done for them?
AS: Imagine that you are tall, work at an office with a professional dress code, have a limited budget, hate synthetic and flimsy materials, and are looking for a new pair of jeans. Normally, you would have to go to dozens of sites, order and bunch and return almost everything or waste hours at the mall sorting through racks of clothes that probably won’t fit you or won’t be quite right. Online stores don’t have the filters for what you want, so you have to keep scrolling.
StyleHacks, on the other hand, helps you find exactly what you want in a matter of minutes, in your size and in your budget. You just tell it what you are looking for (“I need non-stretchy jeans to wear to the office, no synthetic”), give feedback on the items it shows you like you would with a very helpful store associate (“I like these but I’d prefer them a bit darker”), and find the right pair quickly and easily.