How AI is reinventing beauty diagnosis: from dermatological expertise to personalized prescriptions
How to create a dialogue between artificial intelligence and dermatological expertise? That is the challenge an international cosmetics group took on alongside Ekimetrics with the creation of an AI capable of analyzing skin, understanding its needs, and recommending a personalized skincare routine. A pioneering project, now deployed in over 50 countries.
When diagnosis alone is no longer enough
When the group acquired a start-up expert in face analysis, their ambition was clear: to provide a skin diagnosis that is reliable, automated, and accessible to everyone, thanks to AI. But a diagnosis, no matter how precise, is not enough: without interpretation or advice, it remains a static photograph of skin. The initial technology made an approximate diagnosis, without saying what needed to be done. To move from diagnosis to action, the group called on Ekimetrics, with a shared goal: to transform an analysis algorithm into a genuine smart beauty advisor.
A beauty prescription AI inspired by human expertise
- Why recommendations are not enough
Ekimetrics took a bold gamble: to go beyond simple classic “product recommendation”, toward a “beauty prescription”, thanks to an AI capable of acting and thinking like a dermatologist. Unlike traditional e-commerce recommendation engines (“customers like you bought…”), this approach aims to replicate a professional’s judgement, not proposing a product to buy, but a coherent and reasoned skincare routine, based on the skin’s real needs.
- How Ekimetrics’ AI made this possible
Rather than a “black box” model, Ekimetrics created an interpretable AI, built on:
- clear business rules established with the brands’ dermatologists
- a robust algorithmic matching, based on scores attributed to each product according to skin symptoms
- a routine optimization system to ensure coherence between the prescribed products
Each brand in the portfolio was the subject of intensive workshops. Whole days were spent deciphering the formulas, textures, product interaction, contraindications, and marketing priorities. The experts assessed each product; Ekimetrics translated this expertise into algorithmic logic.
3 levers to transform human expertise into artificial intelligence
1. Knowledge engineering
The teams worked brand by brand to recreate a dermatologist’s expertise: Which ingredients treat which symptoms? Which combinations are prohibited? Which textures complement each other? The challenge was not to model the data but to translate experts’ reasoning into algorithmic logic.
2. Interpretability as imperative
Using opaque models was out of the question. Each recommendation had to be explained to the consumers. For example, suppose the diagnosis reveals a loss of skin elasticity. In that case, the AI recommends a particular cosmetic product that contains a specific active ingredient responding to the need for firmness.
This requirement led Ekimetrics to opt for transparent decision trees rather than deep learning.
3. Alignment between science and business
Personalization must also respect each brand’s identity: price positioning, texture, and sensory experience.
Consequently, Ekimetrics integrated business weightings into the algorithm, ensuring a delicate balance between scientific rigor and marketing coherence, and guaranteeing that each category of consumers discovers an experience that meets their expectations and their brand universe.
A new customer experience and a brand-new acquisition lever
- A widely adopted technology
The project was initially deployed to one of the group’s dermatological skincare brands, then to more than ten brands in over 50 countries. Each brand adapted it to their universe, voice, and business priorities, while keeping the same promise: useful, expert, and credible personalization.
- An enriched and transparent beauty experience
For the consumer, the journey is simple:
- A selfie and a few dermatological-type questions
- A detailed diagnosis (firmness, radiance, hydration…)
- An annotated prescription, product by product, routine by routine
- Sustainable business impact
Rather than a short-term performance tool, the solution developed played a strategic role in creating a new acquisition channel.
Observed benefits:
- a high utilization rate in retail and digital
- an increase in customer satisfaction, linked to the pertinence of the advice
- strengthening of the brands’ scientific credibility
- full internalization of the technology by L’Oréal, guarantee of autonomy and longevity
The tool has thus become a new entry point into the brand ecosystem, where the quality of advice takes priority over instant conversion.
This collaboration illustrates a simple conviction: AI does not replace human expertise, it augments it. When it remains explainable and firmly grounded in the reality of usage, it becomes a lever of trust and shared progress, for brands and their clients.
👉 To understand how luxury and beauty Maisons can steer their transformation with AI, read this article: Why luxury and beauty Maisons need AI to steer their transformation.
