Unilever: BeautyHub PRO AI beauty advisor

Unilever’s BeautyHub PRO analyzes up to 30 visual data points from a selfie and recommends skin and hair care products across Unilever brands. Users were 43% more likely to complete a purchase and had baskets nearly 40% larger than shoppers in other channels.

Summary

Item	Detail
Company	Unilever
Industry	Consumer goods
Use cases	Customer experience, personalized recommendations
Functions	Marketing
Reported result	43% more likely to complete a purchase, baskets nearly 40% larger (both vs. other channels); 80,000+ users in 2023, expected to reach 3 million+ people
Moments rating	S-tier quick win
AI use-case score	4.8

Why it matters

A diagnosis-then-recommendation tool gives a multi-brand group a single front door to its whole portfolio. The reported gains are revenue metrics, but they compare tool users with other shoppers, which matters for how they should be read.

Case details

Background

BeautyHub PRO is an AI selfie tool developed in-house by Unilever. Delivered through a downloadable app, it gives skin and hair care advice and has launched in the Philippines and Thailand. Unilever’s original news page cannot currently be reached directly; the details below come from a TechInformed report of August 7, 2024, which quotes Aaron Rajan, Unilever’s global VP of consumer experience technology.

What they did

Users complete a short questionnaire and upload a selfie. A computer vision model assesses up to 30 visual data points, then recommends products from Unilever brands including Dove, Dermalogica, Paula’s Choice, TRESemmé and Vaseline. The algorithms were trained on de-identified data covering a range of skin tones and hair types. The hair damage model, for example, was trained on more than 10,000 standardized selfies collected with consent plus lab images, and tested on more than 150,000 additional data points. Every AI tool goes through privacy, cybersecurity and local legal review before launch.

Results and data

According to the report, consumers who used BeautyHub PRO had baskets nearly 40% larger than people shopping elsewhere, and were 43% more likely to complete a purchase than those browsing in other channels. More than 80,000 people used it in 2023, and it is expected to reach more than 3 million. Users said the questionnaire was simple and the advice credible.

What to take from it

A cross-brand “diagnose, then recommend” entry point suits groups that own many brands. Training data that spans skin tones, and legal and security review before launch, are just as important for any consumer-facing AI tool.

Limits

The 43% and nearly 40% figures compare tool users with shoppers in other channels. People who choose to use the tool may already be more likely to buy, so the numbers should not be read as lift caused by the tool. The 3 million figure is expected reach, not actual users. Collecting face images in China requires separate consent for sensitive personal information under the Personal Information Protection Law.

Source

Original source: Unilever · How AI-powered ultra-personalized experiences are boosting our beauty brands

Additional source: TechInformed report (2024-08-07)

This is an industry research entry: a publicly documented third-party AI case, not a Moments client case. For projects Moments delivered itself, see client cases. The figures above belong to Unilever’s published case and are not a Moments result or promise.