Skip to main content

​
The concept

Groups are a way to group similar product types. Different product types need different data points (attributes). By grouping product types, you can attach only the relevant attributes and allowed values to that group, so enrichment focuses on what matters. How it works in Cernel:
  1. Choose the group that gathers similar products
  2. Add the attributes that matter for that group (for example, Fit, Length, Material, Color)
  3. Define allowed values for each attribute (for example, Fit → Slim, Regular, Loose)
  4. All products under that group are enriched using the same definitions

​
Why this matters

Defining attributes and values per group:
  • Creates predictable structure the AI can follow
  • Keeps enrichment stable and repeatable over time
  • Reduces rework by aligning similar products to the same standards
Result: a smoother, more reliable enrichment flow across your whole catalog.

​
Groups

Groups are a Cernel feature that automatically organizes products based on rules and conditions you define.

​
How Groups Work

Groups use conditions to automatically include products: Example Group Definitions: 
Luxury Items
└── Condition: price > $500

Red Electronics
└── Conditions: category = "Electronics" AND color = "Red"

Bestsellers
└── Condition: sales_rank < 100
Products that match the conditions automatically belong to the group—no manual assignment needed.

​
Group Structure

Groups form a hierarchical tree structure—think of it like a family tree where parent groups contain child groups. Example structure: 
All Products
└── Upper Clothing
    └── T-shirts
        └── Child T-shirts

​
How Inheritance Works

Inheritance means child groups automatically get the rules from their parent groups. You don’t need to repeat the same conditions at every level. Simple example:
  • Parent group “Upper Clothing”: Has condition product_type = "Upper Clothing"
  • Child group “T-shirts”: Automatically inherits product_type = "Upper Clothing" and adds its own condition subcategory = "T-shirts"
This means “T-shirts” actually applies both conditions: product_type = "Upper Clothing" AND subcategory = "T-shirts"
  • Grandchild group “Child T-shirts”: Inherits both conditions from above and adds age_group = "Child"
So “Child T-shirts” applies: product_type = "Upper Clothing" AND subcategory = "T-shirts" AND age_group = "Child" Why this matters:
  • Less work: You set up conditions once at the parent level
  • Consistency: All child groups automatically follow the parent’s rules
  • Simplicity: Child groups only need to add their specific conditions
Start with broad parent groups (like “Upper Clothing”), then create specific child groups (like “T-shirts”) that inherit and refine the rules.

​
Primary Group

Each product has one primary group selected automatically by Cernel. How Cernel Selects Primary Groups:
  1. Specificity: More specific groups are preferred (e.g., “Men’s Formal Shirts” over “Clothing”)
  2. Hierarchy Depth: Deeper nodes in the tree are considered more specific
  3. Condition Complexity: Groups with more conditions are more specific
  4. Alphabetical: If multiple groups are equally specific, alphabetical order breaks ties
Cernel recalculates primary groups automatically when group conditions or product attributes change.

​
Viewing Groups

Navigate to Groups in the sidebar to see:
  • Group Tree: Hierarchical structure of all groups
  • Conditions: The rules that define each group
  • Product Count: How many products match the conditions
  • Enrichment Status: Whether products in this group have AI content

​
Managing Groups

Groups can be created and configured in Cernel:
1

Create Group

Click “New Group” in the Groups page
2

Name Group

Provide a descriptive name (e.g., “Luxury Handbags”)
3

Set Parent

Choose where in the tree this group belongs (or make it a root group)
4

Define Conditions

Add conditions that products must match:
  • Attribute filters (color = Red, size = Large)
  • Price ranges
  • Custom attribute values
5

Save

Cernel automatically assigns matching products to the group
Groups update dynamically—when product attributes change, group membership recalculates automatically.

​
Group-Specific Operations

Groups enable specialized workflows: Example Workflow:
  1. Create group “Products Needing Enrichment” (description is empty)
  2. Enrich all products in this group
  3. As products get descriptions, they automatically leave the group
  4. Use the group to track enrichment progress

​
Customizing AI by Group

You can customize AI behavior per group: Example:
  • Luxury Group: Formal tone, emphasis on craftsmanship and exclusivity
  • Budget Group: Casual tone, emphasis on value and practicality
  • Kids Group: Fun, simple language with safety highlights
Learn how to customize prompts →

​
Managing Groups at Scale

Before enriching your entire catalog, create a test group (e.g., “Sample Products” with 10-20 items) to validate prompt quality.
Set base prompts at top-level groups, then override for specific sub-groups that need different treatment.
Regularly check that primary groups are assigned correctly—these drive AI context and are critical for quality.

​
Common Issues

​
Group Membership Unexpected

Problem: Product is in a group you didn’t expect. Solution:
  1. View the group’s conditions
  2. Check the product’s attributes to see which condition matched
  3. Either adjust the group conditions or modify the product attributes

​
Too Many Products in Group

Problem: Group has thousands of products, making bulk enrichment slow. Solution:
  1. Create sub-groups to break down the group into smaller segments
  2. Enrich in batches rather than all at once
  3. Filter products by additional criteria before enriching

​
What’s Next?

Creating Groups

Learn how to create and configure groups

Jobs & Monitoring

Track enrichment progress

Bulk Operations

Enrich entire groups at once

Primary Taxonomy

Deep dive into primary group selection
Next: Learn how to create groups or check out Jobs & Monitoring to track enrichment operations.