To main website
  1. Help Center - Channable
  2. Optimize
  3. General

Articles in this section

Creating Smart Attributes in Channable

AI enrichments allow you to automatically generate or refine product attributes, creating Smart Attributes to optimize your data efficiently. For more information, refer to our Smart Attributes guide.

To learn how to set up AI enrichments, visit our step-by-step guide, and for common questions, check out the Smart Attributes FAQ.

Introducing Smart Attributes, our AI-powered feature designed to streamline product optimization and boost channel quality. Powered by a large language model, specifically optimized for Channable, AI enrichments can generate, extract, and refine attributes from your imported data. In this article we'll explain in detail what Smart Attributes are and how they work! 

Table of contents

  1. What are Smart Attributes?
  2. Managing AI enrichments
  3. Configuration of the AI enrichment/AI rule
  4. Settings
  5. Targeting (Optional)
  6. Preview (Optional)
  7. Approving results (Mandatory)
  8. Creating AI enrichments

What are Smart Attributes?

Smart attributes are attributes generated through the use of AI enrichments/AI rules in the Optimize tab of Channable. These enrichments automatically extract or generate product attributes such as color, weight and brand from your imported data, reducing the need for manual attribute creation.

For example, if a marketplace requires a color attribute in a separate field, but the color is only available in the product title, Channable’s AI enrichment can extract that color and create a Smart attribute from it.

Explanation of Key Terms

  • Input Data: The fields in your imported data that the AI uses to generate or optimize the attribute values. For example, product titles or descriptions.
  • AI Enrichment/AI Rule: A rule that is configured to search for specific data in a field and generate a Smart Attribute based on that search. It automatically enriches your data by extracting key information, such as extracting "red" from a product title and storing it as the color attribute.
  • Attribute: A specific characteristic of a product (e.g., color, weight, or brand).
  • Smart attribute: The output generated by an AI enrichment, such as a refined or extracted product attribute.

Where are the AI enrichments placed?

AI enrichments or AI rules are applied after the import process is completed in Channable. This ensures that all necessary data has been successfully imported before the AI generates the Smart Attributes. Once the import is finalized, the AI enrichments process runs, generating or refining attributes based on the rules set.

After the AI enrichments are completed, the generated Smart Attributes are merged with the imported data. This ensures that your product information is enriched and optimized before it’s passed through the rest of your project workflow, making it available for use in feeds, exports, and channels.

Supported attributes

Below, you'll find a list of the attributes currently supported by AI enrichments. This list is regularly updated.

  • Color
  • Weight
  • Quantity
  • Brand
  • Length
  • Width
  • Height
  • Volume
  • Material
  • Keywords

Each attribute is generated from your imported data, and the AI can adjust this data based on the language you’ve chosen.

Managing AI enrichments

To manage or generate Smart Attributes, navigate to the Optimize page in your project. This page provides an overview of your AI enrichments, allowing you to create, reorder, regenerate, or navigate through them.

Configuration of the AI enrichment/AI rule

The pop-up that shows after choosing "New AI enrichment" allows you to set up an AI enrichment. Below, you'll find a quick overview.
  • Title: Give the enrichment a name to identify the task.
  • Which type of attribute do you want to fill in?: Select the attribute (e.g., Color) that you want the AI to enrich. Your choice here lets the AI know what data to search for and fill in.

Settings

In the Settings tab, you can select an existing field to store the generated data or create a new field. Creating a new field is recommended to avoid overwriting existing data.

  • Save it in: You can either select an existing field to store the generated data or create a new field. Creating a new field is recommended to avoid overwriting existing data.
  • Output language: The generated data uses the input data’s language by default, but you can translate the output into several supported languages like English, Dutch, German, French, and more.
  • Where to search: Define where the AI should search for the data. Options include searching all project fields, specific fields, or excluding certain fields.

Note: You can always change the settings above at a later moment.

Targeting (Optional)

With targeting, you can specify which items you'd like to enrich. We recommend limiting the scope to ensure faster results. For instance, you can target a specific category or brand to focus the AI’s efforts on.

Preview (Optional)

Preview the generated data for ten products to review the quality before committing to a full generation.

Approving results (Mandatory)

After AI generates the Smart Attributes, you’ll need to review and approve the results. You can now, among other things, approve the full outcome at once or first select individual attributes for review. Approved attributes will be available for use across your Channable project.

Menu on the Results page

The Results page offers several options for managing the attributes generated by AI enrichments. You'll find a breakdown of each option below.

  • Approve all: Approves all generated attributes in one click, making them available for use in your project.

  • Unapprove all: Reverts all previously approved attributes to an unapproved state, allowing you to review or modify them again.

  • Results filter: Allows you to filter the results based on their approval status.

    • All: Shows all generated attributes, regardless of their status.
    • Not approved: Displays only attributes that haven’t been approved (yet).
    • Approved: Displays attributes that have already been approved.
    • No results: Shows the items for which no results were generated.

  • Auto-approve: Automatically approves all generated attributes without manual review. This is useful if you trust the AI’s results and want to save time.

  • Expand all: Expands each result, allowing you to view every item in more detail.

  • Collapse all: Collapses the expanded items to give you a more compact view of the results.

  • Regenerate: Reruns the AI enrichment process based on the most up-to-date data or changes you’ve made, generating fresh results.

  • Finished: Marks the task as complete, signaling that you’ve reviewed and (un)approved the generated attributes.

  • Results generated: Displays the total number of attributes generated by the AI. For example, if the system generates 95 attributes, it will show “Results generated: 95”.

  • Status updates: Provides real-time information about the process, including timestamps, progress, and any important updates.

mceclip4 (1).jpg

Creating AI enrichments

AI enrichments allow you to automatically generate or refine product attributes using AI. If you want to learn how to create and configure AI enrichments step by step, please refer to our detailed guide on creating AI enrichments, where you’ll find instructions on setting up, selecting attributes, and managing your enrichments for optimal results.