Product Category Enrichment

Automatically generate SEO content for your product categories using AI — including page titles, meta descriptions, keywords, descriptions, and URL slugs.

Overview

Product Category Enrichment helps you build fully optimised category pages without writing anything manually. The AI analyses your product catalog or researches the web, then generates structured SEO content for every category you select.

You define the attributes you want filled in Settings → Enrichment → Category Attributes, and the AI takes care of the rest.

Getting Started

Step 1: Configure Enrichment Attributes

Before running any enrichment, set up the content fields you want the AI to generate.

1. Go to Settings → Enrichment → Category Attributes
2. Click Add Attribute
3. Fill in the details:
   • Name: What the field is called (e.g., "Page Title")
   • Slug: A short identifier — use a recognised SEO slug for automatic formatting (see table below)
   • Type: Data type — text, html, select, multiselect, boolean, or number
   • Prompt: Instructions for the AI — describe exactly what you want it to write

Recognised SEO slugs (automatic formatting applied):

When you use one of these recognised slugs, the AI automatically applies the correct format rules without any extra configuration.

Step 2: (Optional) Link Products to Categories

The AI produces its best results when it knows exactly what products belong in a category. Linking products gives it real, specific context.

To link products one at a time:

1. Open any product in Products → [product name]
2. Find the Category field and select the correct category
3. Save

To link products in bulk:

1. Go to Product Categories
2. Click Actions → Map Products to Categories
3. Upload a CSV file with one product and one category per row

Supported product identifiers per row:

Supported category identifiers per row:

Aim for at least 3–5 products per category for best enrichment quality.

Enriching Categories

Single Category Enrichment

From the Category Detail Page:

1. Go to Product Categories and click on any category
2. Click Enrich Category in the top-right corner
3. A dialog appears — configure your options:
   • AI Model — Budget or Pro (see AI Model Comparison)
   • Attribute Set — All active attributes, or a specific set
   • Use Websearch — Enable if the category has no linked products
4. Click Enrich Category to start

A floating progress button appears at the bottom of the page while the job runs. Click it to go directly to the job detail. The page refreshes automatically and updates attribute values when the job completes.

From the Category Table:

1. Go to Product Categories
2. Click the ⋮ (three-dot) menu on any category row
3. Select Enrich Category
4. Configure options and confirm

Bulk Category Enrichment

Step 1: Select Categories

• Check the boxes next to the categories you want to enrich
• Or use Select All to choose all categories on the page

Step 2: Start Enrichment

• Click the Selected (N) dropdown in the toolbar
• Select Enrich Categories

Step 3: Configure Options

• AI Model — Budget or Pro
• Attribute Set — All active attributes, or a specific set
• Use Websearch — Enable for categories with no linked products
• Include existing attributes — On by default. Turn off to regenerate without keeping previous values

Step 4: Confirm

• Click Enrich Categories to start the job
• All selected categories are processed as a single background job — you can leave the page and come back

Bulk Delete Categories

Step 1: Select Categories

• Check the boxes next to the categories you want to delete
• Or use Select All to choose all categories on the page

Step 2: Open the Action Menu

• Click the Selected (N) dropdown in the toolbar
• Select Delete Selected

Step 3: Confirm Deletion

• A confirmation dialog appears listing how many categories will be deleted
• Click Delete to permanently remove all selected categories

Warning: Deletion is permanent and cannot be undone. Any products linked to a deleted category will lose their category association.

Enrich via CSV Upload

Use this option when you have a list of categories to enrich and want to upload them as a CSV file rather than selecting them individually from the table.

Step 1: Open the Enrich page

Go to Product Categories → Actions → Enrich via CSV, or navigate directly to the Enrich page.

Step 2: Prepare your CSV file

Your CSV must contain at least one of these columns:

You only need one — the system matches by name or slug. Download the sample CSV from the page for the exact format.

name,slug
Running Shoes,running-shoes
Trail Running,trail-running

Requirements:

• File type: .csv only
• Maximum size: 8 MB
• Up to 2,000 categories per job
• The category must already exist in your account — new categories are not created

Step 3: Upload the file

Click Select CSV File or drag and drop your file onto the upload area.

Step 4: Configure options

Step 5: Click Enrich Categories

The system reads every row and tries to match it against your existing categories by name or slug:

• Matched rows — included in the enrichment job
• Unmatched rows — skipped and listed in a warning card on the page

If some rows were not found:

A yellow warning card appears below the upload form showing how many rows were skipped and listing each unmatched value with a not found label:

⚠️ 1 category not found and skipped

These rows from your CSV did not match any existing category in your organisation. The enrichment job was started for all matching rows.

not found category Name

The job still runs for all matched rows. Click View Jobs in the warning card to monitor progress. To retry the skipped rows, fix the values in your CSV (check spelling or use the correct slug) and upload again.

If all rows matched:

You are taken directly to the Jobs page to monitor progress.

How the AI Gets Its Information

The AI uses one of two sources depending on your setup:

Linked Products mode — When products are mapped to a category, the AI reads those products (names, brands, descriptions) and generates content that accurately reflects what you sell. No web search is needed — this is the most accurate and cost-effective mode.

Web Search mode — When a category has no linked products and you enable Use Websearch, the AI searches Google, reads competitor category pages, and uses those results as its reference. The source URLs are recorded and shown on the category detail page after enrichment.

Internal Linking

Once you configure Storefront URLs under Settings → Enrichment → General Settings, category enrichment automatically weaves internal links into the generated HTML description:

• Up to 5 brands — the brands with the most products in that category
• Up to 5 related products — products linked to that category

Brands are ranked by product count within the category, so links point at the brands that actually dominate it. Ranking is deterministic: re-running enrichment on the same category produces the same links.

Links are placed only in html-type attributes. Your page title, meta description, meta keywords, and slug never receive anchors. Anchors are plain <a href="…">Anchor Text</a> — no rel, no target — and any URL the AI invents that wasn't supplied by the system is stripped from the result before it is saved.

Entities are linked only when their URL can be built from your base URL and pattern; anything that can't resolve is left unlinked rather than pointed at a broken URL. Leave the brand or product pattern blank to skip that link type entirely.

See Storefront URLs & internal linking for the full configuration reference.

AI Model Comparison

Budget Model

• Best for: Large batches, routine enrichment runs
• Speed: Fastest
• Credit cost: Lowest
• Quality: Strong for most standard categories

Pro Model

• Best for: High-value or top-level categories, flagship pages
• Speed: Slightly slower
• Credit cost: Higher
• Quality: Best-in-class copy quality and SEO precision

Start with Budget for most categories. Use Pro for your most commercially important categories where the content quality has a direct impact on traffic and conversion.

Viewing Enrichment Results

On the Category Detail Page

1. Go to Product Categories and click on any category
2. Open the Enrichment Attributes tab
3. Every configured attribute shows its current AI-generated value
4. Fields not yet enriched appear as blank

A Reference used card below the attributes shows what the AI consulted:

• Linked products mode: Shows the product cards the AI used as context
• Web search mode: Shows the list of URLs the AI read — each is clickable

On the Jobs Page

1. Go to Product Categories → Actions → View Jobs
2. The jobs table shows file name, model, status, and progress counts
3. Click View on any job to open item-by-item results
4. Click Download CSV to export the enriched data

Job Statuses

Attribute Sets

Attribute Sets let you enrich a subset of your configured attributes in a single job — useful for partial updates like refreshing only SEO titles without touching descriptions.

1. Go to Settings → Enrichment → Category Attribute Sets
2. Click Create Attribute Set
3. Give it a name and select the attributes to include
4. When running an enrichment job, select your set from the Attribute Set dropdown

Credits and Costs

Category enrichment uses credits based on:

• The number of categories in the job
• The AI model selected (Budget or Pro)
• Whether web search is enabled

Web search jobs cost more because the AI makes additional calls to research Google search results and competitor pages before generating content. Linked Products mode is cheaper — no external calls are made.

If your credit balance reaches zero during a job, the job pauses and records an error summary. Top up your credits and start a new job to continue.

See Credits & Billing for more on purchasing and managing credits.

Best Practices

Linking Products First

1. Map products before enriching — linked products are the strongest signal the AI has; it knows exactly what is sold in the category
2. Aim for 3–10 products per category — a mix of brands and product types gives broader context
3. Use the bulk map CSV for large catalogs instead of editing each product individually

Writing Good Prompts

Write the prompt as if you were briefing a copywriter. Be specific about tone, length, and what to include.

Good prompt:

"Write a 200-word HTML category description introducing this category to a first-time buyer. Cover what products belong here, who they are for, and the top two buying criteria. Use <h2> and <ul> tags. Avoid brand names."

Poor prompt:

"Write a description"

Cost Optimisation

• Use Linked Products mode wherever possible — it is cheaper than web search
• Use Attribute Sets — only re-enrich the fields you need to refresh
• Start with Budget model — switch to Pro only for your most important categories

Troubleshooting

"No products linked and websearch is disabled"

The category has no products mapped to it and web search is off — the AI has nothing to work from.

Fix: Map at least one product to the category, or re-run the enrichment with Use Websearch enabled.

Enriched values are blank or empty

The AI could not confidently generate a value — usually because the category lacked enough context.

Fix: Add more linked products, enable web search, rewrite the attribute prompt to be more specific, or switch to the Pro model.

Some rows were skipped after CSV upload

One or more rows in your CSV did not match any existing category by name or slug. The enrichment job still starts for all matched rows — only unmatched rows are skipped.

Fix: The warning card on the Enrich page lists every skipped value. Check the spelling against your category names or slugs in the Product Categories table. If the category does not exist yet, create it first, then re-upload the CSV.

Job stuck in "Processing"

Fix: Refresh the Jobs page — it updates every 15 seconds. If the job has been stuck for more than 10 minutes, click Stop and create a new job.

What This Feature Can and Cannot Do

Can do:

• Generate any attribute you configure — SEO fields, HTML descriptions, select/multiselect values
• Use your actual product catalog as grounding context
• Research your category on Google when web search is enabled
• Process up to 2,000 categories per job
• Automatically enforce SEO character limits and slug formatting
• Add internal links to related brands and products when Storefront URLs are configured

Cannot do:

• Create categories that do not already exist in your account
• Enrich a category with no linked products and web search disabled
• Link to a brand or product whose URL cannot be built from your storefront URL pattern

Next Steps

• Product Enrichment — Enrich individual products with attributes and specs
• Attribute Sets — Create reusable attribute bundles for targeted enrichment runs
• Credits & Billing — Purchase and manage credits
• API Keys — Access the platform programmatically