A new beauty startup, Proven, is using artificial intelligence/ AI to determine the best skincare routines for people based on their skin type and needs, and all of the products will be made in-house.
Amy Yuan, one of Proven’s co-founders, uses her computational physics background to build an “AI engine” that pores through reviews of various skincare products, reports TechCrunch. Proven claims the machine has analyzed 8 million reviews, 20,000 ingredients, and 100,000 products using fraud-detection algorithms to rule out fake reviews. By looking through a massive quantity of data, the machine can then determine trends and general patterns in what products are suitable for particular skin types.
Proven suggests a skincare routine for you via a considerably detailed quiz that asks questions like where you live, whether your skin is oily or dry, and what skincare products you can’t live without. At the end, the results categorize what kind of skin you have and offer Proven’s custom skincare products, such as toners, serums, and a non-optional skin report, all of which you can buy as a bundle through a subscription that charges you $120 every two months. The startup appears to be aimed at women, the largest consumer group for the skincare industry, though the quiz does offer the option to identify as male.
Proven is not the first beauty startup to use big data to support and market its product strategy. HelloAva launched a chatbot last year that recommends skincare products that contain safer ingredients. Function of Beauty, a Y Combinator-backed company, also uses algorithms to make custom shampoos and conditioners. Function of Beauty doesn’t claim to use deep learning and machine learning like Proven does, but it does have a detailed quiz and in-house products.
Skincare can be an even tougher field than hair care. The skincare industry is filled with a myriad of products, and everyone’s skin is different, with different skin conditions like eczema and rosacea. Notably, Proven’s quiz fails to ask whether you have these aforementioned conditions. After I took the Proven quiz, it gave me the result of Trouble Skin, probably because I answered “Not Sure” on a question about skin allergies, and listed mild dryness, acne, and oiliness as problems. AI may be useful for Proven’s chemist to figure out the best formula for common skin problems, but from the user end, it doesn’t yet seem personal enough.
You May Like This