Description
Pimcore Product Enrichment Agent module helps teams add AI-based product enrichment to Pimcore. It turns product data into structured field values while keeping product updates under human control.
This module connects Pimcore with OpenAI, Anthropic, Gemini, and Azure OpenAI through one neutral service layer. You can select the provider and credential that fit each enrichment profile.
Create prompt templates for common tasks such as long descriptions, short descriptions, SEO data, attributes, categories, image alt text, brand voice, and localized copy.
Before a batch run, use Inline Preview to test one product. The preview writes nothing, so teams can tune prompts and field mappings before starting a larger enrichment job.
Batch enrichment runs in the background with live progress, item counts, warnings, logs, and recovery tools. The system supports enriching all products or only products that have changed.
Generated values first go to a review stage. Reviewers can compare current and proposed values, edit them, then approve or reject each change. Only approved values are written to products.
Apart from this, to sync Pimcore products, categories, and variants with commercetools, check out Pimcore Commercetools Connector.
Highlighted Features
Structured Product Enrichment
Generate product field values using defined schemas that match the required Pimcore fields.
Enrichment Profiles and Prompts
Create profiles with product classes, field mappings, credentials, and prompt templates for each enrichment task.
Review Before Product Updates
Review, edit, approve, or reject generated values before they are applied to Pimcore products.
Background Enrichment Jobs
Run large catalog updates as background jobs while tracking progress, logs, warnings, and stopped jobs.
Usage and Budget Controls
Set usage limits for runs, jobs, daily activity, and operators while tracking AI usage through a neutral ledger.
Secure AI Credentials
Store multiple encrypted provider credentials, validate them on save, block duplicates, and enforce one default credential per provider.
Why Do We Need Pimcore Product Enrichment Agent module?
Large Pimcore catalogs often contain fields that need manual writing, cleanup, or completion. Creating this data one product at a time can slow down catalog work.
Default product workflows do not provide a controlled AI layer for generating structured values, testing prompts, reviewing results, and applying approved changes.
Pimcore Product Enrichment Agent adds this workflow without giving an AI model direct control over product records. Generated values stay pending until they are approved.
This approach gives catalog teams more control over AI-generated content while reducing repetitive enrichment work. In real catalog operations, review-based updates also help limit unwanted changes.
Use Cases
-
Product Description Generation
Create long and short product descriptions from existing product data. Review the proposed copy before adding it to the catalog.
-
SEO Content Enrichment
Generate SEO meta content from product information and map the result to selected Pimcore fields.
-
Attribute Extraction
Use AI to identify structured product attributes from existing product content and return values in the required output format.
-
Category Suggestions
Use product information to suggest a suitable category. Teams can review the suggestion before applying it.
-
Image Alt Text
Generate image alt text through supported vision models. If the selected model has no vision support, the task is skipped without an API call.
-
Brand Voice Rewriting
Rewrite product copy to match a defined brand voice while keeping the output inside the configured enrichment workflow.
-
Localized Product Copy
Use translation profiles to send existing content to a neutral translation contract and keep the result within the same review process.
-
Changed Product Enrichment
Run enrichment only for products that have changed. This helps reduce repeated processing across large product catalogs.
AI Provider and Credential Management
This module uses one neutral provider interface for all supported LLM services. Provider-specific code stays inside its own directory, making the architecture easier to extend.
- Support OpenAI, Anthropic, Gemini, and Azure OpenAI providers
- Add multiple credentials for each supported AI provider
- Validate API keys before saving new provider credentials
- Reject duplicate credentials before they are stored
- Encrypt API keys at rest using the libsodium library
- Keep credentials out of API responses and application logs
- Set one default credential for each supported provider
- Add new providers without changing the core interface
Enrichment Profiles and Prompt Templates
Prompt templates define the system prompt, user prompt, output schema, and enrichment task. Profiles connect these templates to a product class, field mappings, and a selected credential.
- Create reusable templates for specific enrichment tasks
- Define neutral schemas for structured AI-generated output
- Bind enrichment profiles to product classes by name
- Map generated values directly to Pimcore product fields
- Select the credential used by each enrichment profile
- Support eight built-in product enrichment tasks
- Reuse profiles for repeat enrichment and catalog jobs
Usage, Budget, and AI Governance
This module provides budget controls for teams that need limits on AI enrichment. Limits can be set at run, job, hard, daily, and operator levels.
- Set usage limits for each product enrichment run
- Set usage limits for individual background enrichment jobs
- Apply hard and daily usage caps for AI enrichment
- Control AI usage limits for each individual operator
- Pause enrichment jobs when quota limits are reached
- Return neutral quota errors when requests are blocked
- Track AI usage in a neutral usage ledger
- View usage and budget data directly in Pimcore Studio
Support
For any query or issue, please create a support ticket here http://webkul.uvdesk.com/
You may also check our quality Magento 2 Extensions.