In mid-June, Google announced that shoppers can virtually “try on” women’s tops from brands across Google, including Anthropologie, Everlane, H&M and Loft. Google’s new “guided refinements tool” (which uses machine learning and visual-matching algorithms) lets shoppers fine-tune choices using inputs like color, style and pattern. Google said it internally developed a new image-based AI model that produces lifelike images. When searchers find a product with a “Try On” badge, they can select a model they deem appropriate and the clothing image will reflect how it drapes, folds, clings, stretches and forms wrinkles and shadows on a diverse set of human figures. Google says it will input more sources of information as people test its “search generative experience,” which will be accessible only through its experimental Search Labs.
PepsiCo reported in May that it has been able to unlock cross-basket insights faster from its loyalty programs using AI technology that transforms documents into structured data. Instead of using a manual process that relied on an outside clearinghouse, PepsiCo began using Veryfi’s platform. “We were able to add user-friendly receipt capture and data extraction to our JOY customer loyalty app, reducing our purchase validation time from seven to 11 days [down] to a few seconds, while unlocking unprecedented cross-basket customer data,” said Edgar Reyes, a director at PepsiCo. The information gleaned can include where and when customers are purchasing, brand affinities across the company’s product line as well as competitors, any price and special-promotion sensitivities, and share of wallet and repeat purchasing trends. Veryfi can extract data from 85 currencies, 39 languages and more than 110 defined fields (vendor, total, purchase order, product name, SKU, etc.).
Mercedes-Benz is adding ChatGPT to its “connected car” MBUX system. Mercedes says integrating artificial intelligence (AI) will make the “Hey Mercedes” voice assistant more intuitive. Noting that most voice assistants are limited to predefined tasks and responses, ChatGPT’s large language model will expand the number and complexity of topics it can handle and enable it to respond with more comprehensive answers. Mercedes is collaborating with Microsoft on this integration, but says it will keep driver data private.
Zenni Optical launched an AI-powered “Image Search” tool on its website and mobile app. It enables customers to upload any photo of a frame — whether they saw it online, in print or snapped it themselves in real life — and locate a match out of the retailer’s more than 3,000 frames. By leveraging AI and machine learning, the company claims a user can match a design within seconds. A “Try-On” capability lets shoppers take a five-second video of their face, turning side to side; the AI then calculates the proper sizing.
In June, VSBLTY announced a partnership with Shelf Nine and Phoenix Vision to build a national in-store media network with active digital signage technology that uses AI. Activations will start in New York and roll out nationally into as many as 4,500 stores in Tampa, Florida; Las Vegas and Hawaii. Its AI-driven software uses facial recognition to identify the age, gender and emotion of shoppers, and the way they dwell or interact with screens and QR codes. The tech is already used in Canada, Mexico and other Latin American countries.
In June, SimplyCodes launched a ChatGPT plugin through OpenAI’s plugin store. After installing the plugin to their ChatGPT account, users can ask it to find coupons at 300,000 retailers. SimplyCodes uses crowdsourcing and blockchain to incentivize shoppers to contribute to its coupon database. Later this year, it will introduce iPhone and Android apps as well as additional AI features.
In May, AI-powered cannabis retail platform Genetica announced seed funding to expand into more dispensaries as well as grocery stores and pharmacies — as many as 350 locations in the United States and internationally over the next quarter. Genetica’s “Flora AI,” the underlying technology, is able to deliver personalized product recommendations by analyzing more than 650 million data points, including in-store inventory, customers’ dietary habits, activity levels, age, medical conditions and desired effects.
New York-based start-up Verneek recently launched One Quin, a proprietary AI shopping platform that can field and answer personalized questions through voice activation or text. The company claims this GenAI platform combs through updated public and proprietary data in real-time, providing a type of AI support unavailable on Alexa, Siri, Google Search or ChatGPT, and that it can fulfill even awkwardly phrased requests, for example: “Umm, where’s that salad dressing that’s sugar free?” Investors include a former Coca-Cola CFO, Beyond Meat’s chairman, and the head of J.P. Morgan Chase AI Research. It’s currently in use in the Sprouts chain.
In June, Newegg Commerce announced the use of AI to launch Home Showcase, an interactive online shopping experience. Newegg uses Stable Diffusion AI to improve product photos and illustrate features for customers, and ChatGPT to create succinct product descriptions on abbreviated quick view pages. Home Showcase is available now on Newegg.com.
