Zolotoe Yabloko Adds AI Assistant That Picks Perfume by Emotion

An AI agent first studies the brand's promotional materials to determine which feelings the maker links to a given scent. An algorithm then analyzes customer reviews and adjusts that initial rating to reflect real-world use. The service matches the profile against the shopper's wishes, and only fragrances that match at least 25% make it into the final selection. For new releases with few reviews, the algorithm also draws on customer opinions about individual notes in other perfumes.

Six questions, up to six scents

Users don't need to know perfume terminology. The assistant runs a short six-question survey:

• whether the scent is for the user or a gift;
• the desired feelings, such as confidence, elegance or relaxation;
• the occasion, special or everyday;
• an associated place, from a seaside to a home library;
• the character of the sillage, described through musical imagery, from a melody in headphones to a front-row concert;
• the budget, in four price tiers from an inexpensive impulse buy to no limit.

Afterwards the service builds an emotional portrait of the selection, with a short description such as "soft, sensual, radiant", and returns up to six fragrances with a match percentage. Each can be opened for details, saved to favorites or added straight to the cart.

A wider retail trend

Retailers are adding AI to their services at a growing pace. In January 2026, Yandex Market's Ultima users gained AI-powered virtual clothing try-on from a full-length photo. Lamoda launched a set of digital styling tools, including a chatbot stylist, online fitting and photo search, while 2GIS expanded its assistant to answer questions about public transport.