AI Product Content for Health and Beauty Retailers: INCI Lists, Skin-Type Claims, and Regulatory Copy at Scale

Health and beauty is one of the most content-intensive retail categories in existence. Every SKU carries multiple data layers that must be accurate, consistent, and compliant: ingredient lists formatted to international nomenclature standards, skin-type suitability claims calibrated to UK advertising rules, fragrance allergen declarations required by law, SPF and UVA protection data, shade matrices with undertone and finish data, and certifications (vegan, cruelty-free, organic) that each carry their own declaration requirements.

A mid-sized health and beauty retailer stocks tens of thousands of SKUs across skincare, haircare, makeup, fragrance, supplements, and personal care. Each sub-category has a different attribute model. Each carries different regulatory obligations. New products arrive constantly from brand partners, each delivering data in their own format and at their own level of completeness.

The content team bottleneck in health and beauty is not a shortage of writers. It is the structural impossibility of applying accurate, compliant, category-specific content to a catalogue of this size and complexity using manual processes. This guide covers what that complexity looks like sub-category by sub-category, and how a configurable AI content schema handles it at scale.

INCI ingredient lists: the regulatory foundation

The International Nomenclature of Cosmetic Ingredients (INCI) system is the basis for ingredient declaration on cosmetics products across the UK and EU. Under UK cosmetics regulations (Regulation 1223/2009, now assimilated into UK law), cosmetic products must declare ingredients using INCI names in descending order of concentration. Ingredients present at less than one per cent may be listed in any order after the ingredients present at one per cent or more.

For product pages, this creates a specific content challenge. Brands supply ingredient data in varying formats: some use INCI names correctly, some use common names or trade names, some provide abbreviated or shortened lists, and some provide the data as a paragraph of flowing text rather than a structured ingredient list. The retailer is responsible for presenting the ingredient list accurately and completely on the product page, regardless of what format the brand supplies.

A configurable AI content schema handles this by treating the INCI list as a structured data field with its own formatting rules. The schema defines how ingredients should be declared (comma-separated, descending order, with water as “Aqua”, fragrance as “Parfum” unless specific allergens require individual declaration) and applies that format consistently across every product in the skincare and cosmetics catalogue.

The business case for getting this right is not just regulatory compliance. Ingredient transparency is an active purchase driver in the skincare market. Buyers searching for “niacinamide serum”, “retinol alternative”, or “paraben-free moisturiser” are using specific ingredient terms as search queries. A product page that contains the INCI list accurately is the product page that captures that search traffic.

Fragrance allergen declarations

Beyond the general INCI requirement, UK cosmetics regulations require that 26 specific fragrance allergens be declared individually by name on the product label (and by extension, the product page) when they are present above specified threshold concentrations: 0.01 per cent in rinse-off products, 0.001 per cent in leave-on products.

The 26 allergens include compounds such as Limonene, Linalool, Citronellol, Geraniol, Eugenol, and Cinnamal. They appear in fragranced products including perfumes, moisturisers, shampoos, body washes, and even some makeup products. A product page that lists only “Parfum” without the individual allergen declarations, where those allergens are present above threshold, is non-compliant.

For a retailer stocking hundreds of fragranced products from multiple brands, maintaining allergen declaration compliance across the full catalogue is a significant operational challenge. A schema that includes a dedicated allergen declaration field, populated from the supplier’s full INCI data, addresses this systematically rather than relying on manual checking of each product individually.

Skin-type claims and the CAP Code

The UK Committee of Advertising Practice (CAP) Code and its broadcast equivalent (BCAP) govern what health and beauty product pages can claim. The distinction that matters most in practice is between cosmetic claims (permitted) and medicinal claims (not permitted without appropriate licensing).

A moisturiser product page can claim: “reduces the appearance of fine lines”, “improves skin’s moisture levels”, “leaves skin feeling softer and smoother”. It cannot claim: “removes wrinkles”, “repairs damaged skin cells”, “treats dry skin condition”. The language is specific and the difference between a compliant and a non-compliant claim is sometimes a single word.

Skin-type suitability claims require similar precision. “Suitable for sensitive skin” is a standard cosmetic claim. “Hypoallergenic” implies a level of testing that must be substantiated. “Dermatologically tested” requires that a test was actually conducted. “For eczema-prone skin” moves toward a medical context and requires careful handling.

A configurable AI content schema for health and beauty includes claims language calibrated to CAP Code compliance. The schema defines permitted claim structures for each sub-category (skincare, haircare, sun protection, colour cosmetics) and generates product descriptions that stay within those structures. Where a supplier’s product data includes non-compliant claim language, the schema flags it for review rather than publishing it.

Sun protection: SPF, UVA, and water resistance data

Sun protection products carry a regulatory and consumer information requirement that is more exacting than most other beauty categories.

A complete sun care product page includes:

SPF value: expressed as a number (15, 30, 50) or 50+ for very high protection. The SPF rating must reflect actual in-vitro or in-vivo testing to ISO 24444.

UVA protection: expressed either as a UVA circle logo (indicating UVA protection at least one third of the labelled SPF, per the EU/UK standard), or as a PA+ rating (a Japanese system also used on some UK products). The distinction matters to buyers who know to look for broad-spectrum protection.

Water resistance: expressed as water resistant (holds SPF for at least 40 minutes of water exposure) or very water resistant (80 minutes). Not all sun protection products are water resistant, and the claim must only appear when substantiated by ISO 16407 testing.

Product format and application: lotion, spray, gel, stick, tinted, mineral (zinc oxide/titanium dioxide based), chemical filter, or hybrid. Each has different application instructions and suitability notes.

For retailers stocking large sun care ranges across multiple brands, each of these data points comes from a different supplier in a different format. A schema with dedicated fields for SPF, UVA standard, water resistance, and product format ensures this data is presented consistently and accurately across the full sun care catalogue.

Shade data in colour cosmetics

Makeup and colour cosmetics present a distinct content challenge: the shade matrix.

A foundation may run across 40 or more shades. Each shade needs: shade name, shade code (where the brand uses one), undertone (warm, cool, or neutral), finish (matte, satin, natural, luminous), and skin tone depth descriptor (fair, light, light-medium, medium, tan, deep). Some brands also declare coverage level (sheer, light, medium, full buildable).

A lipstick range across 30 shades needs: shade name, shade code, finish (cream, matte, satin, metallic, gloss), and a colour family descriptor (nude, pink, red, berry, coral) that enables filtering.

The challenge is that this data arrives from brands in inconsistent formats. Some use their own shade system, some use Pantone or RAL references, some provide only a shade name and a swatch image with no structured data. A retailer publishing shade data as received ends up with an inconsistent, unfiltered catalogue where buyers cannot find shades by undertone or finish, which directly suppresses conversion in a category where shade matching is the primary purchase criterion.

An AI content schema that includes structured shade fields, populated consistently from brand data and product imagery, enables shade filtering and makes the makeup catalogue navigable in the way buyers expect from specialist beauty retailers.

Certifications: vegan, cruelty-free, and organic

The health and beauty market has a significant certification landscape, and each certification has a specific declaration requirement.

Vegan: does not contain animal-derived ingredients and has not been tested on animals. The Vegan Society trademark (the sunflower logo) is the most recognised UK standard. Other standards include PETA-approved vegan.

Cruelty-free: not tested on animals, but may contain animal-derived ingredients (e.g. beeswax, lanolin, honey). The Leaping Bunny programme (Cruelty Free International) is the most rigorous standard, with supply chain audits.

Organic: covers both certified organic ingredients (COSMOS ORGANIC, ECOCERT, BDIH, Soil Association) and products with a minimum certified organic content percentage. A product can be “certified organic” or “contains X per cent certified organic ingredients”: the distinction is significant and must be reflected accurately on the product page.

Dermatologically tested / clinically tested: implies a test was conducted. The product page should ideally note the testing organisation or methodology where this is known.

Publishing certifications incorrectly (claiming a product is vegan when it contains beeswax, or claiming a product is Leaping Bunny certified when it is only brand-declared cruelty-free) is a Trading Standards and ASA issue. A schema that treats certification as a structured field with defined permitted values (rather than free text that can be entered inconsistently) reduces the risk of misleading claims at catalogue scale.

The catalogue velocity problem

Health and beauty has high SKU velocity. Major brands reformulate products, launch seasonal editions, introduce new shades, and discontinue lines continuously. A retailer stocking ten major skincare and makeup brands manages a catalogue that is in constant flux.

Every new product, reformulation, or new shade variant needs a complete product page before it can go live. Every discontinued product needs its page updated. Every reformulation may require an INCI list update and a regulatory claims review.

At this velocity, manual content management is not a productivity constraint. It is a structural impossibility. AI product content processes new health and beauty SKUs in batch, applying the correct schema for each sub-category (skincare, colour cosmetics, fragrance, supplements) automatically. New stock is online on the day it arrives, with accurate INCI lists, compliant claims language, and complete shade data.

Internal links

• Configurable AI product content schema: how the attribute model adapts to different product categories
• What is AI product attribution?: how AI extracts structured attributes from product data and images
• AI product content for wholesale distributors: for beauty wholesalers and distributors supplying retailers
• AI retail merchandising hub: platform overview