How to Get Beauty and Skincare Products Recommended by AI Assistants

Beauty is one of the categories where shoppers already rely on AI assistants most. The questions are personal and specific: “best vitamin C serum for sensitive skin under $40”, “a fragrance-free moisturiser safe to use with tretinoin”, “a long-wear foundation for oily skin with a neutral-olive undertone”. Each one forces the assistant to match a product to a skin type, a concern and a constraint. It names only the products it can justify.

How assistants decide which beauty products to name

An assistant answering a skincare question looks for evidence on four points:

• Formulation: does the product contain the ingredients the question implies, at a meaningful level?
• Suitability: is it stated to be suitable for this skin type or concern, and free of what the shopper wants to avoid (fragrance, alcohol, specific allergens)?
• Proof: is there a test, a certification, a dermatologist recommendation or a body of reviews behind the claims?
• Consensus: do independent sources — editorial roundups, retailer reviews, communities — agree?

Your product page can settle the first three. The fourth comes from the wider web, and it is where most brands lose to category leaders.

Step 1. Make ingredients machine-readable

• Publish the full INCI list as text on every product page. An ingredient list that exists only in an image of the box is invisible to a crawler.
• State key actives and their concentrations where you can: “10% niacinamide + 1% zinc”, “0.5% encapsulated retinol”, “SPF 50 with 20% zinc oxide”.
• List what the product is free from in plain words: fragrance, essential oils, alcohol denat., silicones — but only if accurate.
• Explain why each key ingredient is there in one sentence. Assistants often paraphrase this reasoning in their answers.

Step 2. Match the product to skin types and concerns explicitly

Write the matching for the assistant instead of leaving it to infer. A short “Best for” and “Not for” block works well:

• Best for: oily and combination skin, visible pores, post-blemish marks.
• Use with care: very sensitive skin; introduce gradually.
• Do not combine with: other exfoliating acids in the same routine.

For makeup, describe shades in words — undertone (cool, neutral, warm, olive), depth, and finish — and state the coverage and wear time. Shade names like “Biscuit 220” mean nothing to a language model.

Step 3. Back every claim

Beauty claims sit close to health claims, and assistants handle them carefully. For each claim on the page, say where it comes from:

• “Dermatologist-tested” — by whom, on how many people, for what (irritation, sensitisation).
• “Non-comedogenic” — the test or the formulation rationale.
• “Clinically shown to…” — the study size, duration and measured result.
• “Vegan”, “cruelty-free” — the certifying body (for example Leaping Bunny or The Vegan Society).

Vague superlatives (“miracle”, “transformative”, “instant results”) give an assistant nothing to repeat and can reduce trust.

Step 4. Explain how to use the product

Routine questions are among the most common in skincare: “can I use vitamin C and retinol together?”, “where does this go in my routine?”. Give the step in the routine, morning or evening, frequency, amount and what to pair it with. A brand whose pages answer routine questions clearly becomes the source assistants quote for those questions.

Step 5. Open the door to AI crawlers

All of this is wasted if crawlers cannot read the page. Check that AI search crawlers are not disallowed in robots.txt, that your firewall does not block them, and that the product description, ingredient list and price are in the HTML your server returns rather than injected by JavaScript. Add Product and Offer structured data to every product page. The AI-ready product page checklist explains each check.

Step 6. Win the third-party sources

In beauty, assistants lean on sources outside your site even more than in fashion:

• Retailers — Sephora, Ulta, Boots, Space NK, Mecca, Cult Beauty and similar. Their product pages carry ingredients, reviews and Q&A in a consistent structure. Keep your listings complete and identical in substance to your own pages.
• Editorial roundups — beauty publications' “best serums”, “best sunscreens” lists are cited constantly.
• Dermatologist and expert content — articles, videos and posts where professionals explain why they recommend a product.
• Communities — skincare forums and Reddit communities where people discuss results and reactions.
• Reviews — volume matters, and so do specific reviews that mention skin type and results.

Step 7. Measure by question, not by brand

Track a fixed set of questions — by concern, ingredient, skin type and price — across ChatGPT, Gemini, Claude and Perplexity every week. For each question, note who is named and which sources are cited. When a competitor wins, the cited sources show you where the gap is: a missing fact on your page, a retailer listing without your ingredient list, a roundup you are not in.