How merchi.ai Adapts to Any Retail Schema: A Guide to Custom Product Attribute Configuration
Every retail category has a different product data model. Tiles need slip rating, frost resistance, and rectified status. Fashion needs fabric composition, fit, care instructions, and sustainability certifications. Flooring needs wear layer thickness, installation method, and underfloor heating compatibility. Workwear needs EN standards, protection ratings, and certification bodies.
No single fixed template can generate accurate, structured product content across all of these. That is the problem merchi.ai was built to solve.
A health and beauty retailer gave us one of the most demanding schema tests in merchi.ai’s history. Health and beauty is one of the most attribute-dense categories in retail: each product needs skin type suitability, key ingredients, free-from claims, SPF rating, benefit claims, volume, and application method populated accurately and consistently, and that varies significantly between a serum, a cleanser, a day cream, and a targeted treatment. We cleared the product setup backlog using merchi.ai’s configurable schema framework in a single sprint. The platform is a National AI Awards 2026 Finalist (AI SME Business of the Year).
That result would not have been possible with a rigid, preset template. It required a schema that matched the retailer’s exact product data model. Here is how that works.
What a product attribute schema is
A product attribute schema is the structured list of fields that defines what information a product description should contain. It is the data model for a product type.
For a floor tile, that model might include: dimensions (length x width), material (porcelain, ceramic, natural stone), finish (matt, polished, lapatto), slip rating (R9 through R13), frost resistance, wall suitability, floor suitability, and recommended room type.
For a running shoe, the model is entirely different: upper material, midsole technology, drop height (mm), intended surface, stack height, width options, and waterproofing.
In merchi.ai, a schema block defines the attribute model for a product type. You specify the fields, their data types (text, number, select, multi-select, boolean), and any value constraints. When merchi.ai generates content for a product, it populates every field in the schema from the source materials: product images, spec sheets, supplier data, or a combination.
Full documentation is in the schema help guide.
Why one schema does not fit all retail categories
The variation between retail categories is not superficial. The attributes that matter for flooring are structurally different from the attributes that matter for fashion, which are structurally different again from those that matter for tiles or outdoor equipment.
Health & Beauty (the proof case covered above): product type (serum, moisturiser, cleanser, toner, SPF, treatment), skin type suitability (normal, dry, oily, combination, sensitive), key ingredients (with concentration where relevant), free-from claims (paraben-free, sulfate-free, fragrance-free, vegan, cruelty-free), SPF rating, benefit claims (anti-ageing, brightening, hydration, acne control), volume (ml), and application method.
Fashion: fabric composition (percentage breakdown of fibres), fit type (slim, regular, relaxed, oversized), care instructions, colour variant, sustainability certification (GOTS, OEKO-TEX, Responsible Wool Standard), size guide, season, country of manufacture.
Tiles: dimensions (W x H in mm), material (porcelain, ceramic, natural stone), finish (matt, polished, lapatto, textured), slip rating (R9 to R13 for wet areas), frost resistance, wall suitability, floor suitability, room recommendation, rectified edge (yes/no).
A generic AI product content tool that applies the same output template to all three of these will generate accurate content for none of them. merchi.ai’s schema-driven approach generates content that populates the exact fields that matter for each product type, in each retailer’s specific model.
How merchi.ai schema configuration works
The retailer defines the attribute model. merchi.ai generates content that populates it.
Configuration takes place at the product type level. For each product type in your catalogue, you create a schema block that specifies:
- The fields to populate (e.g. “wear layer”, “installation method”, “AC rating”)
- The data type for each field (text, number, select list, multi-select, boolean)
- Value constraints where relevant (e.g. a select field for AC rating with permitted values: AC1, AC2, AC3, AC4, AC5)
- The writing style and length for the description field
- Any brand voice instructions specific to this product type
Once the schema is configured, every product processed against it will have every field populated consistently. The schema enforces the output structure. There is no drift between products, no missing fields, no inconsistent terminology.
Advanced configuration options (writing knowledge, tone of voice templates, and multi-schema workflows) are covered in advanced writing assets and the writing knowledge guide.
What schema-driven generation looks like in practice: the health & beauty example
Health and beauty is one of the most demanding test cases for AI product content. The attribute model is deep, ingredients terminology is technical and regulated, and consistency across hundreds of SKUs from dozens of brands is non-negotiable for both SEO and customer trust.
This is the schema merchi.ai used for that retailer’s catalogue:
| Attribute | Type | Example value |
|---|---|---|
| Product type | Select | Serum |
| Skin type | Multi-select | Oily, combination |
| Key ingredients | Text | Niacinamide 10%, Zinc 1% |
| Free-from claims | Multi-select | Paraben-free, fragrance-free |
| Benefit claims | Multi-select | Pore-minimising, oil control |
| SPF rating | Number | 30 |
| Volume | Number (ml) | 30 |
| Application method | Select | AM and PM, after cleansing |
| Vegan | Boolean | Yes |
| Cruelty-free | Boolean | Yes |
Every product populated consistently to this schema. Attribute values extracted from product images, supplier data, and ingredient lists. No manual data entry. No copywriter. No agency.
The output was not just usable. It was accurate enough to go live immediately, with each product page carrying the complete, structured attribute data that drives both long-tail SEO and on-site filtering performance.
Beyond health & beauty: schema-agnostic by design
merchi.ai is not a health and beauty tool with a nice case study. The schema framework was designed from the start to be category-agnostic. Any product type, any retailer’s data model, any attribute structure.
The platform has generated product content for health and beauty, fashion, flooring, outdoor equipment, tiles, tools, home improvement, and wholesale distribution. Each category uses a different schema. The generation quality does not degrade with complexity: a schema with 20 attributes generates as consistently as one with 5.
For context on how this integrates into the broader retail technology architecture, see where merchi.ai fits in your retail tech stack.
For vertical-specific examples:
- AI product content for fashion retailers covers how merchi.ai handles fabric composition, care, sustainability certifications, and fit across large fashion catalogues.
- AI product content for home improvement retailers covers tiles, tools, fixtures, and the multi-category complexity that home improvement catalogues present.
The automated product classification article covers taxonomy and category classification, which works in conjunction with the schema layer.
The AI retail merchandising platform overview covers the full platform capability.
Configure your schema in 30 days, free
merchi.ai offers a 30-day free trial with full schema configuration access. You define your attribute model on day one. By the end of the first session, you can have a batch of your own products generated against your own schema.
Frequently asked questions
Can AI generate product content for a custom product attribute schema?
Yes. merchi.ai is fully schema-agnostic. You define the attribute model for each product type, including the fields, data types, and value constraints. merchi.ai generates content that populates every field in your schema consistently across your entire catalogue.
What is a product attribute schema in ecommerce?
A product attribute schema is the structured list of fields that defines what information a product page should contain. For health and beauty it might include skin type suitability, key ingredients, free-from claims, and SPF rating. For fashion it might include fabric composition, fit type, and care instructions. The schema varies by product type and by retailer. In merchi.ai, schema blocks define this attribute model and enforce consistent output across every product.
How does merchi.ai’s schema configuration work?
You create a schema block for each product type. You specify the fields, data types, and any value constraints. merchi.ai then generates content that populates every field in that schema from your source materials (images, spec sheets, supplier data). The schema enforces structure and consistency across every product in every batch.
Can I define my own product attributes in merchi.ai?
Yes. There are no preset attribute templates that you are locked into. You define the fields that matter for your products and your customers, set the data types, and specify permitted values where relevant. The platform generates content to your specification.
How does merchi.ai handle different schemas for different product categories?
You create a separate schema block for each product type or category. A retailer selling both flooring and wall tiles might have two distinct schema blocks with different fields. Products are processed against the appropriate schema based on their category. There is no limit on the number of schema blocks.
Does merchi.ai work with ETIM, GS1, or Shopify Taxonomy?
Yes. merchi.ai supports standard classification frameworks including ETIM, GS1, and Shopify Taxonomy as taxonomy outputs alongside custom schemas. You can use a standard framework as the classification layer while maintaining a custom attribute schema for your content generation. See the automated product classification article for detail.
How does schema configuration affect SEO for product pages?
A complete, consistent attribute schema produces product pages that match the structured queries buyers use in search (“niacinamide serum for oily skin”, “paraben-free SPF 30 moisturiser”). Completeness and attribute specificity are direct inputs to product page SEO performance. Retailers using merchi.ai typically see improved long-tail search rankings as a direct result of schema completeness across their catalogue.