Yes — an independent retailer can offer a conversational skin assessment on their own site, without the budget or the technical teams of a major brand. The AI skin tools built by L'Oréal, La Roche-Posay and Vichy are proprietary: they run on their sites, to sell their products. They don't help you sell yours. The accessible alternative is a conversation that does an in-store advisor's job: it asks a few questions about the customer's skin and habits, then recommends only from your catalogue, never outside it. It is never a medical diagnosis — it is a conversational routine assessment, which points to products and hands over to your team the moment a sensitive case appears (pregnancy, allergy, a complaint). Ryma, the AI-powered advisor, already works this way at a cosmetics client in production.
Why the big brands built their own AI skin tool
The major brands have each developed their own AI skin-assessment tool, built straight into their sites: L'Oréal, La Roche-Posay and Vichy are good examples. These tools read the visitor's skin and then recommend a product — but only one kind of product: theirs.
From their point of view that is entirely logical: they have a proprietary catalogue to move, a dedicated technical budget, and a direct interest in steering every assessment towards their own range. The problem is that those tools do nothing for you as a reseller or an independent brand. Even if you stock those products, the brand's tool will never recommend your store: it pushes the sale to them. You can't lean on it to sell from your own catalogue, and you can't build an equivalent either without resources few independent retailers have.
That is the real subject when you look at these tools and wonder « how do I get the same thing »: the question isn't how to copy the tool, but how to reproduce what it does for the customer — on your catalogue.
The alternative: a conversation that does an in-store advisor's job
The alternative available to an independent retailer is an assessment in the form of a conversation, not a proprietary scientific instrument. In practice the advisor asks the customer a few questions: skin type, main concern, current routine, products already used. Then she recommends, based on the answers, from the store's catalogue. Never from outside it.
That is the essential difference from the big brands' tools: here the assessment serves your stock, not a third party's. Her work doesn't stop at the answer — it ends at the cart. The role isn't to answer a technical question and move on, but to lead the customer to the right product, exactly as an advisor in a physical shop would.
To place this approach against a plain e-commerce chatbot, the distinction between an AI agent and a chatbot is useful: a chatbot answers questions, an agent advises and actively recommends.
What such an assessment's credibility rests on
The credibility of a conversational skin assessment rests on three non-negotiable pillars, and on a serious volume of testing before it goes into production. First pillar: the advisor is fed the official documentation of the major brands. Second pillar: she works from a register of guidance reviewed and approved by health professionals (a dermatologist, a paediatrician, a pharmacist). Third pillar, and the most important one for trust: on sensitive cases — pregnancy, allergy, a complaint — she hands over to the retailer's team, by a fixed rule.
These pillars are not an abstract statement of intent. Before going into production at Beautymarket, the number one beauty and skincare store in Morocco, the Ryma advisor went through 100,000 test conversations. The catalogue in question is dermo-cosmetic: L'Oréal, La Roche-Posay, Vichy, Avène, Bioderma, CeraVe. It is that upstream testing, on a demanding catalogue, that lets a conversational assessment hold up without derailing on the cases where caution has to win.
What it actually changes for your sales
A conversational skin assessment changes the sale by replacing a silent product page with a guided recommendation — and that shows up in measured results at several brands. At bareMinerals, a cosmetics brand, a case study published by Gorgias measured a return of 8.8 times the investment in the first month. At Topicals, a skincare brand, the same source measured sales driven by customer advice growing 78%.
There is a reason underneath those figures: in beauty, people repurchase a product that kept its promise, far more than one that was cheap, according to McKinsey's State of Beauty 2025. A well-run assessment points the customer to the product most likely to keep that promise for them, which feeds that repurchase directly. That is the whole point of the Beautymarket case, where the mechanism runs at the scale of a high-traffic site.
At Beautymarket, the Ryma advisor handles 70% of requests end to end with no human involvement (measured in testing), on a site that receives 1,000,000 visitors a month (traffic measured by SimilarWeb).
The limits to know before you start
The first limit, and the most important, is that a conversational skin assessment is never a medical diagnosis: it points to products and advises, and it in no way replaces the opinion of a health professional. That distinction has to stay in mind at every step — not sit as a legal line at the bottom of the page.
The second limit concerns sensitive cases. Pregnancy, a declared allergy, a complaint: on those subjects the rule is fixed, and the advisor hands over to the retailer's team rather than risking an unsuitable recommendation. That is a design choice, not a gap: a fast human relay beats badly calibrated advice on a health question.
The third limit is about data. An assessment means collecting information about the customer, which imposes data minimisation from the design stage: collect only what serves the recommendation, host in Europe, anonymise before the AI models, and put a proper data processing agreement in place. The full detail of those guarantees is on the reliability and GDPR page.
Frequently asked questions
›Is an online skin assessment a medical diagnosis?
No — never. A conversational assessment points the customer towards products suited to what they describe, but it in no way replaces the opinion of a dermatologist or another health professional. On cases that call for particular caution, such as a pregnancy or a declared allergy, the advisor hands over to the retailer's team rather than carrying on alone.
›Does the customer have to send a photo of their skin?
No, no photo is needed. The assessment rests on a conversation: a few questions about skin type, the main concern and the current routine are enough to point to the right products in the catalogue. It is a conversational routine assessment, not image analysis.
›Do my customers have to create an account to do it?
No — the conversation can start without an account, like a chat with an advisor in a shop. That removes friction and lets the customer ask their question or start straight away, on the site, on WhatsApp or on social channels, with the same memory of the catalogue from one channel to the next.
›Which products does the advisor draw on to recommend?
Only your store's catalogue, never products from outside it. Setup starts from the Ryma app authorised inside your Shopify store — or from an export of that catalogue if you are on another platform. A direct Shopify connection can be added afterwards for stock and orders; see the guide on installing an AI on Shopify.
›What if the customer doing the assessment is pregnant?
That is precisely one of the cases covered by the sensitive-case rule: pregnancy always triggers a hand-over to the retailer's team, rather than an automatic product recommendation. It is not a gap in the tool — it is a deliberate choice never to stand in for medical advice on a subject that calls for it.
Alexandre builds Ryma, the AI-powered advisor for online stores, and writes these guides from real deployments, with every figure attributed.