Here, Dr. Svetlana Kordumova, an AI and Computer Vision graduate, talks about founding Pixyle.ai, its AI-powered solutions, and the future of retail. Pixyle.ai aims to leverage AI in e-commerce to optimize product data entry, increase customer engagement, and enhance the shopping experience.
Focusing on the fashion industry, Kordumova envisions a future where AI reshapes the way we shop and operate e-commerce platforms.
WWD: What was the impetus for founding your company? What problems did you think needed solving in the market?
Svetlana Kordumova: I was a Ph.D. student at the University of Amsterdam, specializing in AI and computer vision. On a personal note, I am also an avid online shopper. The shopping experience I encountered on fashion e-commerce platforms was frustrating. It was difficult to find what I wanted. At the same time, it was also working on advanced AI technologies with image recognition and search capabilities, which had not yet been applied to real user scenarios in e-commerce.
I still remember attending an event and the presenter was wearing a dress that perfectly matched what I had imagined. I couldn't help but wonder why there wasn't a way to find that dress online just by taking a photo. So I think my frustration with the challenges of online shopping and my experience with AI and image understanding inspired me to start her Pixyle.ai.
The biggest challenge I saw in the marketplace was making it easy for shoppers to find what they wanted. Pixyle.ai's core product is focused on using AI to solve manual data entry problems for e-commerce teams, and this is still part of our core value. Our mission is to provide e-commerce teams with superior AI product data to save time and ultimately help shoppers find the products they love, effortlessly.
WWD: What other challenges are facing retailers and brands in terms of consumer engagement?
SK: In today's world of diverse technologies and AI tools, retailers and brands still face the challenge of effectively engaging consumers. One big problem is that e-commerce sites have poor filtering or search that doesn't work well, so shoppers struggle to find what they want. This makes it difficult for consumers to find products they like online, resulting in them abandoning purchases and e-commerce stores losing out on sales.
All of this involves the dreaded process of manually entering product data by your e-commerce or merchandising team. So, to optimize time and process, enter only basic data and tags such as item category and color. However, if a person wants to buy “Puff Sleeve Dress for Beach Wedding” and types it into the search bar, the e-commerce store does not have the sleeve type “Puff Sleeve” as data associated with that dress. If you use a “beach wedding” as an occasion, your customers won't be able to see the results. He has 50 of these dresses in his e-commerce store, but since the dresses are not visible to customers with high purchase intent, they may never reach a sale.
Data associated with products is critical to many touchpoints in the customer experience. All tools like personalization, recommendations, and search work better when data is richer. Marketing teams can run more targeted ads, and SEO teams can use their SEO-optimized product tags, titles, and descriptions to improve rankings. Decision makers will have richer and better data on which to base future collection and merchandising decisions.
You can see that the fashion industry is still very traditional in terms of operations. Teams still do a lot of things manually. I believe AI has the potential to change this by automating many processes. Pixyle.ai automates complete product data entry for your e-commerce team. Using his Pixyle.ai, his clients can get product tags, attributes, titles, descriptions, site search, SEO-optimized tags in multiple languages from a single product image in seconds. Masu.
WWD: How do AI solutions work?
SK: Pixyle.ai is a product data enrichment platform that uses computer vision AI and generative AI to generate product data for fashion e-commerce. Pixyle's AI models can automatically tag and organize your online store's product information in multiple languages, saving your e-commerce team time on manual tasks.
Svetlana Kordumova
Brands use our AI platform to solve the pain of having to manually allocate product data, which takes too much time and resources. This manual process is time consuming and inaccurate. Inaccurate product data makes it difficult for shoppers to find your products. Insufficient data also limits online product filtering options. This inevitably reduces sales and conversions.
The challenge for marketplaces is manually processing data for large volumes of products from different brands, which is time-consuming and costly. Extensive teams collaborate on product data entry procedures. To save time, this manual process is limited to only a few attributes, so it can be used to filter products online or provide relevant search results when you type a query into the search bar. Not enough data available.
Powered by computer vision AI, our product tagging solution saves e-commerce teams valuable time by automatically tagging images with rich attributes. For example, for a dress sold on an e-commerce store, the attributes that Pixyle detects are “red,” “long sleeves,” “floral,” “V-neck,” “business casual” style, and “Pixyle's AI detects , all this data can be generated in just 0.9 seconds by processing a single product image of a dress. If this data had to be entered from scratch by a human, it would take on average about 3 minutes per product. Masu.
Simply put, our visual AI engine generates tags from photos uploaded by users or clients, depending on the business model of your e-commerce store. This allows you to automatically tag your entire catalog in minutes, eliminating the need to hire or train a dedicated team, helping you meet deadlines and improve product discoverability.
Pixyle.ai also provides optimized site search tags. This enriches your inventory with accurate and rich AI-generated tags from a dictionary of over 20,000 attributes collected from users' search queries. This improves the relevance of site search results to e-commerce stores and simplifies shopping. For example, if a customer searches for “beach wedding party puff sleeve dresses” on an e-commerce site with strong purchase intent, advanced search functionality won't matter if your product doesn't have any relevant sleeve types or occasions. yeah. Even if the store has more than 100 such dresses for him in stock, they will not appear in the results and you will miss the opportunity to show high sales motivation.
Label recognition is another AI solution offered by Pixyle.ai that extracts clothing label data such as the exact brand, size, origin, and material of a fashion item from neck and side labels. This solution is used in the e-commerce market of the second-hand goods industry. By taking a photo of your label and sending it to Pixyle's label recognition AI, you can now enter all the information on your label in seconds without having to manually enter your precious time.
Label recognition is like having a fashion-savvy friend who can instantly decipher clothing labels. Imagine you sell on a second-hand goods platform or work at a vintage clothing thrift store. Instead of squinting and filling out tiny text, just take a photo of the label and you're done. Pixyle's AI works its magic and provides all the information for your label in seconds. No more tedious data entry.
WWD: What is the value proposition for the retailers and brands you use? What makes you different from other AI solutions on the market?
SK: We help e-commerce teams save time entering and enriching product data. We've seen up to an 80% increase in efficiency for merchandising and e-commerce teams. Additionally, after using Pixyle's AI data enhancements, our clients measured significant increases in sales and conversion rates. With better product information, shoppers can more easily find what they want, leading to more sales. Conversion rates increased by 8% and average order value increased by 35%.
Pixyle stands out from its competitors because it is the only platform on the market dedicated to data entry for fashion. All our taxonomies, AI models, and innovations are tailored specifically for the fashion industry. This effort enables us to offer the richest taxonomy available, ensuring high accuracy and relevance in all aspects of product tagging and data enrichment. With a focus on fashion, we have refined our tools to meet industry-specific needs and nuances, establishing ourselves as the go-to solution for fashion e-commerce businesses.
WWD: How do you see the use of generative AI and other technologies evolving in the coming years?
SK: In the future, I think we'll see generative AI and other technologies move beyond the general hype and into more niche areas to solve specific problems. AI can help fashion brands design clothes, predict trends, and even customize outfits for customers. This means there may be more unique and cool clothing options to suit people's tastes. It will shake up the way fashion works, making it faster, more creative, and better suited to what people want. The entire shopping experience changes from a search to a conversation. When you visit an e-commerce store, a digital agent will guide you through the virtual store, just like in a physical store.
I believe this will revolutionize both the way we shop and the way e-commerce teams operate. Enhanced intelligence allows e-commerce teams to innovate faster and deliver better, more personalized experiences to shoppers.
