Skip to main content

What are AI Workflows

AI Workflows allow you to automate repetitive catalog tasks in Elovate using AI powered actions and rule based triggers.

Instead of manually running separate enrichment, translation or validation jobs, AI Workflows let you build automated pipelines that process products continuously or in batches.

A workflow starts with a trigger. Once products match the configured rules, one or multiple AI actions are executed automatically.

Examples of actions include:

  • Attribute Extraction

  • Content Enrichment

  • Translation

  • Category Mapping

  • Content Validation

  • Quality Scoring

This makes it possible to automate large parts of your product data operations directly inside Elovate.


How AI Workflows work

An AI Workflow consists of two main parts:

Trigger

The trigger determines which products should enter the workflow.

You can define this using rules and filters such as:

  • Category equals Shoes

  • Description is empty

  • Status is enabled

  • Brand equals Nike

This ensures only relevant products are processed.


Actions

Actions are the AI powered steps executed after a trigger is activated.

Actions run sequentially, creating a waterfall effect where one action can continue based on the output of a previous action.

Example workflow:

  1. Extract missing attributes

  2. Generate optimized shopping descriptions

  3. Translate content to German

  4. Validate content quality

  5. Synchronize data to your platform

This allows you to fully automate enrichment pipelines.


Why use AI Workflows?

AI Workflows help you:

  • Reduce repetitive manual work

  • Automate product enrichment

  • Generate structured attribute data

  • Improve catalog consistency

  • Scale multilingual content

  • Build automated publishing pipelines

  • Improve data quality control

  • Manage AI generation in a single overview

Instead of managing disconnected AI jobs, workflows centralize your automation into one scalable system.


Workflow execution types

AI Workflows currently support different execution methods.

One-time workflows

A one-time workflow processes the products that currently match your rules once.

This is useful for:

  • Initial catalog enrichment

  • Bulk translations

  • Large cleanup projects

  • Supplier data enrichment


Continuous workflows

Continuous workflows continuously monitor your catalog and automatically process new matching products over time.

This is useful for:

  • Automatically enriching new products

  • Ongoing translation pipelines

  • Automated quality control

  • Continuous synchronization flows


Workflow moderation

AI Workflows include moderation capabilities to help maintain content quality.

Depending on your configuration:

  • editors can review generated output

  • moderators can approve changes

  • AI reasoning can be inspected

  • confidence scores can be reviewed before synchronization

This gives teams more control over AI generated data before publishing.


AI Workflows replace old enrichment jobs

Previous Content Enrichment and Translation jobs are migrated into the new AI Workflow system.

The new workflow system provides:

  • a centralized overview

  • chained AI actions

  • clearer execution visibility

  • moderation improvements

  • scalable automation pipelines

Existing jobs are automatically converted into workflows where possible.


Common AI Workflow use cases

AI Workflows can be used for many catalog automation scenarios.

Examples include:

  • Extracting product attributes from descriptions

  • Generating Google Shopping content

  • Translating product content automatically

  • Validating missing catalog data

  • Assigning categories with AI

  • Creating multilingual enrichment pipelines

  • Improving incomplete supplier content

Did this answer your question?