# Welcome to the fabric Knowledge Base
Source: https://developer.fabric.inc/home
fabric’s API-first platform gives developers the control and flexibility to easily build from scratch or integrate with the existing stack, to compose the perfect commerce experiences.
Welcome to the fabric NEON Knowledge Base
Explore how to use NEON to monitor product visibility and enrich product data for AI search.
Looking for documentation on fabric’s legacy products?
See the
v3
or
v2
docs.
Learn how Product Agent can enrich your product data to fabric NEON's golden standard.
Measure your catalog’s visibility across AI search and agentic commerce prompts.
Enrich product data for AI search and publish it across your sales channels.
***
# Experiments
Source: https://developer.fabric.inc/product-agent/activate/activating-products/experiments
Test how enriched product content impacts performance.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
An **Experiment** is a controlled A/B test that compares how enriched products (variant) perform against their original versions (control) over time.
Experiments measure how enriched product content impacts performance before you roll changes out across your full catalog. This allows you to test a subset of products, compare results, and apply data-driven decisions at scale.
Experiments are **optional** but recommended.
After you enrich and publish the products in the Experiment in Activate, performance data is tracked in [Monitor](/product-agent/monitor/experiment-performance).
## Why Run Experiments
Experiments help you validate the impact of enrichment by:
* Confirming that enriched content improves visibility and engagement
* Measuring performance changes using real data
* Reducing risk before scaling updates
* Building confidence in your content strategy
## Procedures
Experiments use the standard Activate workflow. The only difference is how you define your product set at the start.
To create an experiment:
1. In the left menu, click **Product Agent** > **Activate**.
The **Product Activation** page is displayed. It shows all previously uploaded products. First-time users will not see any products.
2. Click **Create Experiment** and do one of the following:
* To create an experiment using new products, click **Via CSV Import**, then follow the steps in [Importing Products](/product-agent/activate/activating-products/importing-products).
* To create an experiment using existing products, click **From Existing Products**, then follow the steps in [Enriching Products](/product-agent/activate/enriching-products).
Experiments begin after enriched products are published. To complete an experiment, continue through the remaining steps in Activate:
* Review the [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping).
* Review the [Sampled Enrichment](/product-agent/activate/sampled-enrichment).
* Complete [Enrichment](/product-agent/activate/enrich-all-review).
* [Publish](/product-agent/activate/publishing-products) the enriched products.
* [Track performance](/product-agent/monitor/experiment-performance) in Monitor using the experiment analytics.
# Importing Products
Source: https://developer.fabric.inc/product-agent/activate/activating-products/importing-products
Import product data through CSV upload to initiate enrichment.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The Product Activation page gives you visibility into your product data and category structure that Product Agent uploaded and ingested. This catalog forms the foundation for enrichment and scoring.
You can use this page to:
* Browse the products and categories that have been uploaded.
* Search and sort by product name, category, or ingestion status.
* Confirm that your core product data is ready for enrichment.
## Prerequisites
Provide your fabric Delivery Manager with two CSV files: one for your categories and one for your attributes. Your delivery manager will coordinate with you to ensure these files are uploaded and mapped.
## Product Information File Guidelines
Product information is uploaded through a CSV import. fabric recommends that you use the provided CSV template as a basis for your upload. It contains predefined column headers, some of which, such as SKU, category, breadcrumb, title, description, and images are required. However, the CSV template is set up for apparel, so there are a number of `attribute*` columns that won’t apply to every merchant.
The `attribute*` columns can be edited to fit the needs of your brand. For example, a furniture manufacturer would need `attribute*height`, `attribute*length`, and `attribute*width`.
The following table describes each column header included in the CSV template:
| **Column** | **Required** | **Description** |
| :------------------------------- | :----------------- | :------------------------------------------------------------------------------------------------------- |
| **sku** | Yes | Unique product ID. This must match your internal SKU and remain consistent across updates. |
| **category** | Yes | Primary product category. Used for benchmarking and enrichment. |
| **breadcrumb** | Yes | Full category hierarchy for the product, separated by a delimiter. Example: Home > Living Room > Chairs. |
| **title** | Yes | Product name as displayed on your PDP. |
| **description** | Yes | Full product description from your PDP. |
| **images** | Yes | URLs to product images. Multiple URLs can be separated by commas. |
| **attribute.colors** | Category dependent | Product colors. |
| **attribute.sizes** | Category dependent | Available sizes. |
| **attribute.materials** | Category dependent | Primary materials used in the product. |
| **attribute.age\_suitability** | Category dependent | Intended age group or suitability range. |
| **attribute.brands** | Recommended | Brand name associated with the product. |
| **attribute.patterns** | Category dependent | Pattern or print description. |
| **attribute.gender** | Category dependent | Intended gender audience. |
| **attribute.closure\_types** | Category dependent | Closure mechanism, such as, zipper, button, snap. |
| **attribute.care\_instructions** | Category dependent | Care or maintenance instructions. |
| **attribute.sleeve\_lengths** | Category dependent | Sleeve length information. |
## Procedure
1. In the left menu, click **Product Agent** > **Activate**.
2. Do one of the following:
* To import products, click **Import**.
* To create an experiment using a CSV file, click **Create Experiment**, then select **Via CSV Import**.
The **Upload CSV** menu is displayed.
If your products are already activated, you can skip this step by clicking **Enrich Products** and following the steps in [Enrich Products](/product-agent/activate/enriching-products).
3. (Optional) Click **Sample CSV** to download a sample file.
The file is downloaded.
4. To import a CSV file, do one of the following:
* Drag and drop a file into the upload box.
* Click anywhere inside the upload box to select a file from your computer.
5. Click **Start Processing**.
Product Agent processes your upload. Depending on the size of the file, it may take some time.
Once your products have been processed, you are automatically redirected to the next step in the process, [Taxonomy Mapping](/product-agent/activate/activating-products/taxonomy-mapping.mdx).
### Replacing or removing products
If you need to update or remove products after an import, upload a new CSV file with the corrected data.
To replace product information, include the same SKUs with updated values in the new CSV file. Product Agent will process the file and apply the updated data during import.
To remove products, delete those SKUs from the new CSV file before uploading.
To upload a new file, follow the steps above in **Importing Products**.
## Related Topics
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping)
* [Enriching Products](/product-agent/activate/enriching-products)
* [Sampled Enrichment](/product-agent/activate/sampled-enrichment)
* [Enrich All & Review](/product-agent/activate/enrich-all-review)
* [Publishing Products](/product-agent/activate/publishing-products)
# Introduction
Source: https://developer.fabric.inc/product-agent/activate/activating-products/introduction
Define your product set and see how taxonomy mapping prepares your catalog for enichment.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
Product Activation is the first step in Activate. In this step, you define the set of products you want to work with.
You can either:
* Import new products using a CSV file
* Select products that are already available in your catalog
This creates the working set of products that you’ll enrich, review, and publish.
### Experiments
When defining your product set, you can choose to work with your full catalog or create an Experiment using a subset of products.
Experiments let you test how a smaller group of products performs after enrichment before applying changes across your full catalog.
For more information, see [Experiments](/product-agent/activate/activating-products/experiments).
## Defining Your Catalog
To begin, bring products into Activate by [importing them from a CSV file](/product-agent/activate/activating-products/importing-products).
After your products are available in Activate, you’ll continue to enrichment to refine and optimize your product data.
## Taxonomy Mapping
Once your products are available in Activate, Product Agent automatically classifies them using its taxonomy mapping system.
This helps Product Agent understand what types of products you’re working with and apply the appropriate enrichment logic.
For example:
* A clothing brand might include shirts, pants, and jackets
* A home goods store might include coffee makers, chairs, and kitchen tools
This classification enables Product Agent to:
* Apply the correct enrichment logic, such as sizing for apparel or dimensions for appliances
* Determine which product attributes are required or recommended
* Tailor downstream steps based on product type
For best results, define your catalog structure in [Taxonomy](/product-agent/admin-settings/settings/taxonomy-overview) before activating products. This ensures Product Agent uses your categories and attributes as the source of truth during classification and enrichment.
### Manual mapping
In most cases, Product Agent classifies products accurately without manual input.
If your data includes custom categories or unclear labels, you may need to:
* Confirm or refine suggested classifications
* Assign more specific product types
* Reorganize products that were misclassified
## Related Topics
* [Activate Overview](/product-agent/activate/overview)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
* [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping)
* [Enriching Products](/product-agent/activate/enriching-products)
* [Sampled Enrichment](/product-agent/activate/sampled-enrichment)
* [Enrich All & Review](/product-agent/activate/enrich-all-review)
* [Publishing Products](/product-agent/activate/publishing-products)
# Enrich All & Review
Source: https://developer.fabric.inc/product-agent/activate/enrich-all-review
Browse enriched content and compare Product Agent's enrichment to your original data.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
After reviewing the sample enrichment, the Enrich All & Review page gives you a final opportunity to assess the content generated by Product Agent before applying it across your catalog.
In this step, you can browse through all enriched products, compare AI-generated content with your original data, and make edits or additions as needed. Your feedback helps Product Agent learn your tone, correct misunderstandings, and generate more brand-accurate content moving forward.
After enriching your products, Product Agent displays a list of your products with the following information:
| **Column** | **Description** |
| :------------------- | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Product** | Shows the product name and thumbnail image to help you identify each enriched product in your catalog. |
| **Attributes Added** | The number of new or optimized attributes Product Agent created for the product. Attributes include details such as fit, material, care instructions, colors, length, and other structured data used in AI answers and search experiences. |
| **FAQs Added** | The number of new AI-generated FAQs added to the product. These improve AEO performance by giving AI engines structured, question/answer content. |
| **Actions** | Details of Product Agent’s optimizations for the specific product. |
## Procedure
1. Click **Enrich Now**.
Product Agent enriches your catalog. When the process is finished, a list of the enriched products is displayed, along with metrics tracking Product Agent’s changes.
2. Review Product Agent’s optimizations of your catalog.
3. Click **Next: Publish**.
You are directed to the next step in the process, [Publishing Products](/v3/product-agent/activation/publishing-products).
## Related Topics
## Related Topics
* [Activate Overview](/product-agent/activate/overview)
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
* [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping)
* [Enriching Products](/product-agent/activate/enriching-products)
* [Sampled Enrichment](/product-agent/activate/sampled-enrichment)
* [Publishing Products](/product-agent/activate/publishing-products)
# Enriching Products
Source: https://developer.fabric.inc/product-agent/activate/enriching-products
Select products or categories to enrich and begin the enrichment process.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
After reviewing the categories and products that Product Agent activated, the next step in the process is enriching your products.
When you click **Enrich**, Product Agent analyzes your product data and generates optimized descriptions, attributes, FAQs, metadata, and schema based on your source information and brand guidelines.
After enrichment is complete, you can review and edit the generated content in the Sampled Enrichment screen before publishing.
You can choose to enrich all your products, or only products in specific categories.
## Procedure
1. On the **Product Enrichment** screen, do one of the following:
* To enrich products belonging only to certain categories, click **Categories**.
* Select the categories you want to enrich.
* To enrich all products, click **All products**.
2. Click **Next: Sampled Enrichment**.
You are directed to the next step in the process, [Sampled Enrichment](/product-agent/activate/sampled-enrichment).
## Related Topics
* [Activate Overview](/product-agent/activate/overview)
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
* [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping)
* [Sampled Enrichment](/product-agent/activate/sampled-enrichment)
* [Enrich All & Review](/product-agent/activate/enrich-all-review)
* [Publishing Products](/product-agent/activate/publishing-products)
# Overview
Source: https://developer.fabric.inc/product-agent/activate/overview
Prepare and enrich your product catalog for AI-driven search.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Activate** process generates AI-optimized product data, including titles, descriptions, attributes, FAQs, and image metadata. This process ensures your product detail pages (PDPs) are structured so that AI-powered shopping assistants and recommendation engines discover them.
The Activate workflow is structured, iterative, and guided. You can preview AI-generated content before it’s applied, provide feedback to fine-tune results, and export your enriched catalog for downstream systems.
Whether you’re preparing for a full-scale syndication strategy or looking to optimize high-priority categories, Product Agent gives you the tools to enrich at scale.
### Experiments
Activate lets you create [Experiments](/product-agent/activate/activating-products/experiments) to test how a subset of products performs after enrichment and validate results before applying changes across your full catalog.
## Activation Workflow
The process of activating your products follows these steps:
1. [Product Activation](/product-agent/activate/activating-products/introduction): Import products or select existing products to define the catalog you want to work with. You can also choose a subset of products to run an experiment.
2. [Enriching Products](/product-agent/activate/enriching-products): Choose the products or categories from your catalog to enrich.
3. [Sampled Enrichment](/product-agent/activate/sampled-enrichment): Review AI-generated sample content for a subset of products. Compare optimized and original data, make edits, and train the system with your feedback.
4. [Enrich All & Review](/product-agent/activate/enrich-all-review): Apply enrichment across your selected products and evaluate the changes using a summary table showing scores, improvements, and AI-generated content.
5. [Publishing Products](/product-agent/activate/publishing-products): Export your enriched catalog as a CSV, ready to ingest into your ecommerce platform or syndication pipeline.
Throughout the process, Product Agent learns from your preferences and continuously improves its output to reflect your brand voice, product structure, and content standards.
# Publishing Products
Source: https://developer.fabric.inc/product-agent/activate/publishing-products
Publish your enriched products to your ecommerce platforms.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
After enrichment is complete, **Publish** is the final step in the Activate process. Publishing pushes your enriched product data to your connected sales channels.
If you began the enrichment process by creating an [experiment](/product-agent/activate/activating-products/experiments), it becomes active after publishing. You can view the results on the [Performance](/product-agent/monitor/experiment-performance.mdx) page in Monitor.
### Exporting enriched data
In addition to publishing, this step also allows you to export data in the following formats:
* **CSV**: Export a spreadsheet of your enriched product data.
* **JSON-LD**: Export enriched, structured [schema.org](https://schema.org/) data for your products and FAQs.
Use the CSV export to review enriched data and import into your internal systems. Use the JSON-LD export options to manually implement structured product and FAQ data on your PDPs.
## Procedure
1. (Optional) To download your enriched data, click one or more of the following:
* Export CSV
* Export Product Schema (JSON-LD)
* Export FAQ Schema (JSON-LD)
The file is downloaded.
2. Click **Publish to Channels**.
Your enriched product data is published across your connected ecommerce channels.
## Related Topics
* [Activate Overview](/product-agent/activate/overview)
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
* [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping)
* [Enriching Products](/product-agent/activate/enriching-products)
* [Sampled Enrichment](/product-agent/activate/sampled-enrichment)
* [Enrich All & Review](/product-agent/activate/enrich-all-review)
# Sampled Enrichment
Source: https://developer.fabric.inc/product-agent/activate/sampled-enrichment
Review and refine AI-generated product content before applying enrichment across your catalog.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
**Sampled Enrichment** allows you to review and refine AI-generated product data before applying it to your catalog.
In this step, Product Agent displays the attributes, FAQs, image alt text, and metadata for the product scope you selected in the previous step. You can review the changes Product Agent made, see its rationale, make edits, and provide guidance to improve future enrichment results.
## Page Layout
The Sampled Enrichment screen is divided into two main areas:
* Product list (left panel): Displays all products included in enrichment. You can filter by category and select individual products to review.
* Product detail (main panel): Displays the enriched data for the selected product.
Selecting a product from the left panel displays that product’s enriched data in the main panel.
Each product includes the following sections detailing its data:
| **Section** | **Description** |
| :----------------- | :---------------------------------------------------------------------------------------------------------- |
| **Attributes** | The product description and details such as size, color, material, or instructions. |
| **FAQs** | Questions and answers designed to improve product discovery and support search-based experiences. |
| **Image alt tags** | Descriptions for product images to improve accessibility and image search visibility. |
| **Meta data** | SEO-focused fields such as meta title and meta description to improve search engine visibility and ranking. |
### Attribute sources
Attributes are grouped into two sections to help you understand where each value comes from:
* **From Your Taxonomy**: These attributes come from your existing product data. Product Agent may standardize or enhance these values to improve consistency and clarity.
* **From Neon Data Set**: These are attributes that Product Agent identifid as opportunities based on fabric’s data model and industry benchmarks. They represent important product details that may be missing from your current data but can improve discoverability, filtering, and AI readiness.
### Enrichment status
Each piece of product data includes a status indicator to show how it was handled during enrichment:
* **New**: Newly generated by Product Agent
* **Updated**: Modified from the source data
* **Original**: Unchanged from the source data
These statuses help you quickly identify which content was created, enhanced, or left as-is.
### Enrichment insights
Hovering over a piece of product data displays additional context, including:
* Agent rationale explaining why a value was generated, updated, or left unchanged
* Sources used to derive the value
* Attribute origin (for example, supplier data or AI-generated content)
This information helps you understand how Product Agent determined each value.
Hovering over attributes that Product Agent enhanced from your taxonomy gives you the **Revert** button so that you can undo the enrichment.
## Procedure
1. Navigate through the sampled enrichment of your catalog and review Product Agent’s optimizations.
2. (Optional) To make changes to Product Agent's optimizations or the guidance that Product Agent uses when enriching your content, see [Optional Procedures](/product-agent/activate/sampled-enrichment#optional-procedures) below.
3. Click **Next: Enrich All & Review**.
You are directed to the next step in the process, [Enrich All & Review](/product-agent/activate/enrich-all-review).
## Optional Procedures
The following procedures explain how to adjust optimizations in Sampled Enrichment, whether by editing a single product or providing guidance for future enrichments.
Your edits and feedback help Product Agent learn your preferences and generate better results.
### Regenerating AI content
The product description, attributes, FAQs, and meta data all support AI edits at the individual level. This allows you to use the Product Agent AI to regenerate content without manual edits or rewrites.
1. Mouse over a piece of product data and click the magic wand icon.
The guidance window is displayed.
2. In the text field, give the AI guidance on what you would like changed.
3. Click **Generate**.
Product Agent regenerates the data.
You can regenerate data as needed.
### Editing a product's title
1. To change a product's title, mouse over the title and click the pencil icon.
The **Edit Title** window is displayed.
2. In the **Title** field, enter the new title.
3. In the **Provide edit guidance** field, provide a reason for the edit.
4. In the **Scope** field, choose whether the edit applies to:
* Only this product
* Every product with this attribute
5. Click **Save**.
The new product title and guidance are saved.
### Editing an attribute
1. Mouse over the attribute and click the pencil icon.
The edit menu is displayed.
2. Edit the attribute as required.
3. In the **Scope** field, choose whether the edit applies to:
* Only this product
* Every product with this attribute
4. In the **Guidance for the agent...** field, provide a reason for the edit.
5. Click **Save**.
The edits and guidance are saved.
### Editing an FAQ
1. Mouse over the FAQ and click the pencil icon.
The edit menu is displayed.
2. In the **Question** field, edit the question.
3. In the **Answer** field, edit the answer.
4. In the **Add AI guidance for handling similar changes...** field, provide a reason for the edit.
5. Click the checkmark icon.
The edits and guidance are saved.
### Editing image alt tags
1. Mouse over an alt tag and click the pencil icon.
The **Edit Alt Text** window is displayed.
2. In the **Alt Text** field, edit the alt tag as required.
3. In the **AI guidance for handling similar changes** field, provide a reason for the edit.
4. Click **Done**.
The edits and guidance are saved.
### Editing meta data
1. Mouse over the meta data and click the pencil icon.
The edit menu is displayed.
2. Edit the attribute as required.
3. In the **Meta Title** field, edit the meta data.
4. In the **Guidance for the agent...** field, provide a reason for the edit.
5. Click the checkmark icon.
The edits and guidance are saved.
## Related Topics
* [Activate Overview](/product-agent/activate/overview)
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
* [Taxonomy Mapping](/product-agent/activate/taxonomy-mapping)
* [Enriching Products](/product-agent/activate/enriching-products)
* [Enrich All & Review](/product-agent/activate/enrich-all-review)
* [Publishing Products](/product-agent/activate/publishing-products)
# Taxonomy Mapping
Source: https://developer.fabric.inc/product-agent/activate/taxonomy-mapping
Map product categories and attributes to your brand's structure.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
This topic covers how to review and adjust category and attribute mappings after you upload your products.
When processing the CSV file of your products, Product Agent attempts to match your source categories and attributes to its internal catalog structure. Classifying your products during this step ensures the right enrichment logic is applied to each product, for example, distinguishing between size, capacity, or material depending on the product type.
In most cases, mappings are handled automatically and require no action. However, it is always a best practice to review the mappings for accuracy.
Taxonomy Mapping is divided into two tabs: **Categories** and **Attributes**. Both tabs allow you to review the mappings that Product Agent applied and change them as you require.
The Categories tab allows you to hover over an individual product and click **View** to see its entire attribute set imported from your CSV file.
Both tabs feature dropdown menus so that you can make adjustments.
fabric recommends that you complete your brand's master taxonomy mapping in **Product Agent** > **Settings** > **Taxonomy** prior to uploading products to Activate. Completing this process helps Product Agent create a structured catalog that supports consistent, high-quality product enrichment. For more information, see [Taxonomy Overview](/product-agent/admin-settings/settings/taxonomy-overview).
## Prerequisites
Ensure you have completed the previous step in the Activate process, [Importing Products](/product-agent/activate/activating-products/importing-products).
## Primary Procedure
1. Review the **Categories** and **Attributes** that Product Agent identified to ensure they are correct.
If you need to make changes to how Product Agent identified your **Categories** or \*\*Attributes, see the [Secondary Procedures](/product-agent/activate/activating-products/taxonomy-mapping#secondary-procedures) section below.
2. Click **Import and Enrich**.
The **Optimizing a Sample of Your Products** window is displayed.
3. Click **See Enrichment Sample**.
You are redirected to the next step in the process, [Sampled Enrichment](/product-agent/activate/sampled-enrichment).
## Secondary Procedures
### Changing a product's category
1. From the **Categories** tab, browse your products.
2. For any mis-categorized product, click on its **Category**.
The category selector menu is displayed.
3. Select the matching category.
The product's category is updated.
### Adding a new category
1. From the **Categories** tab, browse your products.
2. For any mis-categorized product, click on its **Category**.
The category selector menu is displayed.
3. Click **Create new category**.
The *Create New Category*\* window is displayed.
4. In the **Category Name** field, enter a name.
5. (Optional) In the **Parent Category** field, select this new category’s parent category.
6. Click **Create**.
The new category is created. Follow the steps above under [Changing a product's category](/product-agent/activate/activating-products/taxonomy-mapping#changing-a-product’s-category) to assign the product to the new category.
### Changing an attribute's mapping
1. From the **Categories** tab, click the **Attributes** tab and browse the attributes.
2. For any mis-mapped attribute, click on its mapping.
The attribute selector menu is displayed.
3. Select the matching mapping.
The attribute's mapping is updated.
### Adding a new attribute type
1. From the **Categories** tab, click the **Attributes** tab and browse the attributes.
2. For any mis-mapped attribute, click on its mapping.
The attribute selector menu is displayed.
3. Click **Create new attribute**.
The *Create New Attribute*\* window is displayed.
4. In the **Attribute Name** field, enter a name.
5. In the **Attribute Key** field, enter an identifier for use in your system, such as "color" or "size."
6. (Optional) In the **Description** field, enter a description.
7. Click **Create**.
The new attribute type is created. Follow the steps above under [Changing an attribute's mapping](/product-agent/activate/activating-products/taxonomy-mapping#changing-an-attribute’s-mapping) to update attribute mapping.
## Related Topics
* [Activate Overview](/product-agent/activate/overview)
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
* [Enriching Products](/product-agent/activate/enriching-products)
* [Sampled Enrichment](/product-agent/activate/sampled-enrichment)
* [Enrich All & Review](/product-agent/activate/enrich-all-review)
* [Publishing Products](/product-agent/activate/publishing-products)
# Brands
Source: https://developer.fabric.inc/product-agent/admin-settings/admin/brands
View brands and parent company associations in Product Agent.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Brands** tab in the Admin section of Product Agent is a read-only tab that displays information about all of the brands you created when setting up Users and Companies.
In Product Agent, a **company** represents the top-level organization. This may be a retailer, brand owner, or agency.
A company can manage one or many **brands**. In some cases, a company and brand are the same entity. In others, such as agencies, a single company may manage multiple brands.
The Brands tab displays a table with the following information:
| Column | Description |
| :------- | :------------------------------------------------------------------------------------- |
| Domain | The brand’s URL. |
| Company | The brand’s parent company. |
| Products | The Product Agent module the brand has access to, whether **Monitor** or **Activate**. |
| Created | The date the brand was created. |
## Related Topics
* [Overview](/product-agent/admin-settings/admin/overview)
* [Users](/product-agent/admin-settings/admin/users)
* [Companies](/product-agent/admin-settings/admin/companies)
# Companies
Source: https://developer.fabric.inc/product-agent/admin-settings/admin/companies
Create and manage companies in Product Agent and organize brands under their parent companies.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Companies** tab in the Admin section of Product Agent allows you to create and manage companies within your organization.
In Product Agent, a **company** represents the top-level organization. This may be a retailer, brand owner, or agency.
A company can manage one or many **brands**. In some cases, a company and brand are the same entity. In others, such as agencies, a single company may manage multiple brands.
The Companies tab displays a table with the following information:
| Column | Description |
| :------ | :---------------------------------------- |
| Name | Company name. |
| Brands | The brands associated with the company. |
| Created | The date the company profile was created. |
## Prerequisites
You must have **Admin** credentials to create or manage companies.
## Procedures
### Creating a new company
To create a new company, take the following steps:
1. In the left navigation of Product Agent, click **Admin**.
The Users tab is displayed.
2. Click the **Companies** tab.
The **Companies** tab is displayed.
3. Click Create Company.
The **Create Company** window is displayed.
4. In the **Name** field, enter the company’s name.
5. In the **Brands (optional)** field, assign the company its associated brands.
6. Click **Create Company**.
The company is created.
### Editing a company
To edit an existing company, take the following steps:
1. In the left navigation of Product Agent, click **Admin**.
The Users tab is displayed.
2. Click the **Companies** tab.
The **Companies** tab is displayed.
1. In the list of companies, click on a company’s name.
The **Edit Company** window is displayed.
2. Edit the company's information as required.
3. Click **Save Changes**.
The company is updated.
## Related Topics
* [Overview](/product-agent/admin-settings/admin/overview)
* [Users](/product-agent/admin-settings/admin/users)
* [Brands](/product-agent/admin-settings/admin/brands)
# Overview
Source: https://developer.fabric.inc/product-agent/admin-settings/admin/overview
Manage users, companies, and brands in Product Agent.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Admin** section controls access and organizational structure in Product Agent.
Unlike **Monitor** and **Activate**, which are used to analyze and enrich product content, Admin defines who can use Product Agent and which brands are available for monitoring and activation.
Admin is organized into the following tabs:
Here it is in clean Markdown format, unchanged:
| Tab | Description |
| :-------- | :------------------------------------------------------------------ |
| Users | Create and manage users. |
| Companies | Organize companies (organizations) and group associated brands. |
| Brands | View brand domains and manage their access to Monitor and Activate. |
### Organizational structure in Product Agent
Depending on your business model, a company may manage a single brand or multiple brands. In Product Agent:
* A company represents the parent organization (such as a retailer, brand owner, or agency).
* A company can manage one or multiple brands.
* Brands can be provisioned for Monitor, Activate, or both.
## Related Topics
* [Users](/product-agent/admin-settings/admin/users)
* [Companies](/product-agent/admin-settings/admin/companies)
* [Brands](/product-agent/admin-settings/admin/brands)
# Users
Source: https://developer.fabric.inc/product-agent/admin-settings/admin/users
Create and manage users in Product Agent by assigning roles and organizing users by company.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Users** tab is the default tab under Admin. It allows you to create and manage users in Product Agent. This topic follows the procedures to create and manage users.
The Users tab displays a table with the following information:
| Column | Description |
| :------ | :------------------------------------- |
| Name | User’s name. |
| Email | User’s email address. |
| Company | The company the user works for. |
| Created | The date the user profile was created. |
## Prerequisites
You must have **Admin** credentials to create or manage users.
## Procedures
### Creating a new user
To create a new users, take the following steps:
1. In the left navigation of Product Agent, click **Admin**.
The **Users** tab is displayed.
2. Click **Create User**.
The **Create User** window is displayed.
3. In the **Name** field, enter a name.
4. In the **Email** field, enter the new user's email address.
5. In the **Company** field, click the company the user works for.
6. In the **Role** field, chose the user's role from the following privilege levels:
* **User** to access and use Product Agent.
* **Admin** to manage users and administrative settings, as well as all privileges the User role has.
7. In the **Brands (optional)** field, assign the user to the specific brands they work on.
8. Click **Create User**.
The user is created.
### Editing a user
To edit an existing user, take the following steps:
1. In the left navigation of Product Agent, click **Admin**.
The **Users** tab is displayed.
2. In the list of users, click on a user’s name.
The **Edit User** window is displayed.
3. Edit the user’s information as required.
4. Click **Save Changes**.
The user is updated.
## Related Topics
* [Overview](/product-agent/admin-settings/admin/overview)
* [Companies](/product-agent/admin-settings/admin/companies)
* [Brands](/product-agent/admin-settings/admin/brands)
# Overview
Source: https://developer.fabric.inc/product-agent/admin-settings/settings/brand-config-overview
Configure the guidelines that Product Agent uses when generating enriched product content.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
With the **Brand Configuration** tab, you can control how Product Agent generates enriched content for your products. You can set the voice, languages, channels, and competitors to ensure AI enrichment aligns with your standards and performance goals.
**Brand Configuration** is broken down into the following sections:
## Channels to Optimize Content For
The **Channels to Optimize Content For** menu allows you to select the platforms where you want your enriched product data to be published. Each channel has its own formatting, tone, and metadata requirements. For example, **Amazon Marketplace** has its own product listing requirements that are different from those of **TikTok Shop**.
You can select multiple channels and Product Agent will create tailored content for each.
## Languages to Generate
The **Languages to Generate** menu lets you choose which languages Product Agent generates enhanced product data in. English is the default. You can also choose to have content enriched in French, Spanish, or German.
## Data Enhancement Options
The **Data Enhancement Options** menu allows for optional improvement of product content by crawling your product suppliers. When enabled, Product Agent retrieves data from supplier websites to use as additional information when enriching content.
This may increase processing time and token usage. It’s meant for deeper enrichment when product data is sparse or inconsistent.
## Competitor URLs
Use the **Competitor URLs** menu to Link Product Agent to up to five of your competitor’s websites. This helps Product Agent in the enrichment process by providing:
* Content benchmarking
* Attribute comparison
* Market positioning analysis
* Industry standards
Brand Guidelines
The **Brand Guidelines** menu allows you to define the voice, tone, and writing style that Product Agent uses when enriching content. This setting ensures that AI-generated product descriptions, headlines, SEO text, and FAQs align with the brand's identity, whether your tone is polished and elegant, casual and friendly, or fast-paced and promotional.
Content tone influences how customers perceive your brand, how easily product details are understood, and how well your listings perform across search engines and marketplaces. Configuring a consistent tone helps build trust, brand recall, and conversion performance.
Carefully crafting your brand tone gives you:
* **Consistency across channels**: Maintains a unified voice whether content is surfaced on Chat GPT, Perplexity, or your own PDP.
* **Customer alignment**: Tailors language to your ideal audience, whether that’s professionals, value shoppers, or eco-conscious buyers.
* **Automation clarity**: Sets clear guardrails when rewriting or enriching copy so the results don’t feel robotic or off-brand.
## Related Topics
* [Monitor Tuning](/product-agent/admin-settings/settings/monitor-tuning)
* [Brand Config Setup](/product-agent/admin-settings/settings/brand-config-setup)
# Brand Config Setup
Source: https://developer.fabric.inc/product-agent/admin-settings/settings/brand-config-setup
Steps to configure channels, languages, competitor inputs, supplier data enhancement, and brand voice settings in Product Agent.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
This topic covers the steps to set up Brand Configuration.
## Procedure
1. In the left menu, click **Settings**.
The Product Agent Settings menu is displayed.
2. Click **Brand Configuration**.
The **Brand Configuration** tab is displayed.
3. In the **Channels to Optimize Content For** section, select one or more channels that you want Product Agent to create enriched product data for.
4. In the **Languages to Generate** section, select one or more languages to generate content in.
English is the default. French, Spanish, and German are also supported.
5. (Optional) To have Product Agent crawl your suppliers’ websites to gather additional information about your products for enrichment, in the **Data Enhancement Options** section, click **Crawl Product Suppliers**.
6. (Optional) To have Product Agent benchmark your enriched content against your competitors’ content, in the **Competitor URLs** field, do the following:
* In the **Competitor 1 URL** field, enter a competitor’s URL.
* To add another competitor, click Add Competitor URL and enter a competitor’s URL.
You may enter up to five competitor URLs.
7. In the **Brand Guidelines** section, configure your global guidelines by doing the following:
* In the **Brand Tone & Voice Guidelines** field, enter a new guidance rule.
* Click **+ Add**.
The new brand tone and voice guideline is added.
Repeat this step as required to create additional rules.
Configuration is complete. You can return at any time to adjust your setup.
## Related Topics
* [Monitor Tuning](/product-agent/admin-settings/settings/monitor-tuning)
* [Brand Config Overview](/product-agent/admin-settings/settings/brand-config-overview)
# Creating Attributes
Source: https://developer.fabric.inc/product-agent/admin-settings/settings/creating-attributes
Define product attributes and data types for consistent enrichment.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
Defining **Attributes** helps Product Agent understand what types of information exist in your catalog and how that information should be interpreted during enrichment. A well-defined set of attributes improves consistency across your product data and gives Product Agent better context when generating enhanced content.
This topic describes how to import attributes from a CSV file or create them manually.
### CSV file guidelines
If you import attributes using a CSV file, the file should define the structure of your product attributes in a clear, consistent format.
Each row represents a single attribute, including its display name, a machine-readable key, and the type of data it stores. This allows Product Agent to understand how product information should be structured and interpreted during enrichment.
At a minimum, attribute definitions typically include:
* **attribute\_name**: the display name shown in Product Agent
* **attribute\_key**: a standardized, machine-readable identifier
* **description**: a brief explanation of the attribute
* **data\_type**: the type of value the attribute stores, such as text or number
Providing a consistent set of attributes ensures that Product Agent can apply enrichment reliably across your catalog.
The following example shows how attribute data can be structured in a CSV file:
| **attribute\_name** | **attribute\_key** | **description** | **data\_type** |
| :------------------ | :----------------- | :------------------ | :------------- |
| Designer | designer | Product designer | STRING |
| Product Name | product\_name | Name of the product | STRING |
| Size | size | Product size | STRING |
| Product Code | product\_code | Internal SKU/code | NUMBER |
| Details | details | Product details | STRING |
## Procedures
### Importing attributes
1. In the left menu, click **Settings**.
The Product Agent **Settings** menu is displayed.
2. Click **Taxonomy**.
The **Taxonomy** tab is displayed.
3. Click **Attributes**.
The **Attributes** menu is displayed.
4. To import attributes from a CSV file, click **Import Attributes**.
The **Import Attributes** window is displayed.
5. To import a CSV file, do one of the following:
* Drag and drop a file into the upload box.
* Click **Choose File** to select a file from your computer.
6. Click **Import**.
The imported attributes are added to your taxonomy and become part of the data structure Product Agent uses during enrichment.
### Manual attribute setup
1. In the left menu, click **Settings**.
The Product Agent **Settings** menu is displayed.
2. Click **Taxonomy**.
The **Taxonomy** tab is displayed.
3. Click **Attributes**.
The **Attributes** menu is displayed.
4. To create a new attribute, click **Create New Attribute**.
The **Create New Attribute** window is displayed.
5. In the **Attribute name** field, enter the name of the attribute.
6. (Optional) In the **Description** field, enter a description for the attribute.
7. In the **Data Type** field, select the type of value the attribute will store.
For more information on data types, see the [Taxonomy Overview](/product-agent/admin-settings/settings/taxonomy-overview).
8. (Optional) To allow Product Agent to enhance or generate values for this attribute, click **Allow AI Content**.
9. (Optional) To provide additional guidance for AI-generated values, click **Include AI Content Guidance**.
10. Click **Create**.
The new attribute is added to your taxonomy and becomes available for Product Agent to use when enriching product data.
## Related Topics
* [Taxonomy Overview](/product-agent/admin-settings/settings/taxonomy-overview)
* [Creating Categories](/product-agent/admin-settings/settings/creating-categories)
* [Importing CSV Files Using API Endpoints](/product-agent/developer-guides/postman-import-csv-guide)
# Creating Categories
Source: https://developer.fabric.inc/product-agent/admin-settings/settings/creating-categories
Create category hierarchies to organize your catalog and improve product enrichment.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
**Categories** define the structure of your catalog in Product Agent.
Creating a clear category hierarchy helps Product Agent understand how your products are grouped during enrichment. A well-structured category tree improves consistency across your catalog and gives Product Agent better context when generating enhanced product content in Activate.
This topic describes how to import categories from a CSV file or create them manually.
### CSV file guidelines
If you import categories using a CSV file, the file should represent your catalog hierarchy in a clear, structured format.
Each row defines a category and its position within your hierarchy using a breadcrumb path, such as `Womens > Apparel > Dresses`. This allows Product Agent to reconstruct your category tree and understand how products are grouped during enrichment.
You can optionally include category identifiers, such as `category_id` or `parent_category_id`, if your catalog is managed in an external system. If these values are not provided, Product Agent automatically generates any necessary identifiers when building your taxonomy.
The following example shows how category data can be structured in a CSV file:
| **\_id** | **category\_name** | **breadcrumb** | **parent\_category\_id** |
| :------- | :----------------- | :------------------------- | :----------------------- |
| | Womens | Womens | |
| | Apparel | Womens > Apparel | |
| | Dresses | Womens > Apparel > Dresses | |
| | Blouses | Womens > Apparel > Blouses | |
| | Skirts | Womens > Apparel > Skirts | |
During enrichment, Product Agent uses your category structure to determine where products belong. Continuing with the example from the table above, if you uploaded a category of t-shirts, Product Agent would assign them to the most relevant existing category, in this case, **Womens > Apparel**.
However, Product Agent does not create new categories automatically. If you want products to be classified under a more specific category such as **Womens > Apparel > T-Shirts**, you must define that category in your taxonomy before enriching your catalog.
## Procedures
### Importing Categories
1. In the left menu, click **Settings**.
The Product Agent **Settings** menu is displayed.
2. Click **Taxonomy**.
The **Taxonomy** tab is displayed.
3. To import your category tree from a CSV file, click **Import Categories**.
The **Import Categories** window is displayed.
4. Do one of the following:
* Drag and drop a file into the upload box.
* Click **Choose File** to select a file from your computer.
5. Click **Import**.
The new category structure is imported and becomes part of the catalog framework Product Agent uses during enrichment.
### Manual category setup
1. In the left menu, click **Settings**.
The Product Agent **Settings** menu is displayed.
2. Click **Taxonomy**.
The **Taxonomy** tab is displayed.
3. To create a category tree manually, do the following:
* Click **New Category Tree**.
The **New Category Tree** window is displayed.
* In the **Category name** field, enter the name for the category.
* (Optional) To add a child or subcategory, click the icon to the right of the **Category name** field.
The **Level 2** field is displayed.
* (Optional) Repeat the above step as required to accurately define your catalog structure.
4. Click **Add**.
The new category structure is added and becomes part of the catalog framework Product Agent uses during enrichment.
## Related Topics
* [Taxonomy Overview](/product-agent/admin-settings/settings/taxonomy-overview)
* [Creating Attributes](/product-agent/admin-settings/settings/creating-attributes)
* [Importing CSV Files Using API Endpoints](/product-agent/developer-guides/postman-import-csv-guide)
# Monitor Tuning
Source: https://developer.fabric.inc/product-agent/admin-settings/settings/monitor-tuning
Configure customer personas, regional targeting, and category price ranges to refine visibility scores in Monitor.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
Use **Monitor Tuning** to refine how Product Agent evaluates your brand’s visibility and relevance.
These settings influence the prompts and scoring logic used during Monitoring. Changes apply to your next Monitor run after saving.
You can tune:
* The Customer persona to shape audience-specific visibility scoring
* Different geographic Regions to limit analysis to specific geographic markets
* The category pricing ranges to align scoring with your pricing strategy
Changes to Monitor Tuning apply to the next monitoring cycle. Previously completed cycles are not updated retroactively.
Monitor Tuning is divided into the following sections:
## Customer Persona
In **Customer Persona**, you define your target customer to guide Monitor’s analysis and scoring. Monitor uses the persona you create here to interpret product relevance, evaluate the audience, and generate more precise visibility scores.
fabric recommends that you be specific when setting up the customer persona. Include demographic traits, style preferences, shopping behavior, and price sensitivity where applicable.
## Regions Selection
Select one or more regions in **Regions Selection** to constrain visibility scoring and competitive evaluation for targeted monitoring in those specific markets.
If no regions are selected, Monitor evaluates visibility without geographic constraints.
## Category Pricing Range Inputs
Use **Category Price Range Inputs** to set competitive value assumptions and improve scoring accuracy. The minimum and maximum fields allow you to tell Product Agent the estimated price range for each category, allowing Product Agent to monitor your content against direct competitors.
For example, a high-end furniture retailer with an “end tables” category priced between $1,000–$2,000 would not typically compete directly with a value-tier retailer offering end tables priced between $100–$200.
## Best Practices
You can return at any time to adjust Monitor Tuning and your changes will be reflected in the next monitoring cycle. fabric recommends that you update Monitor Tuning when you shift your target audience, expand into new regions, or adjust your pricing strategy.
## Procedure
1. Navigate to **Product Agent** > **Settings** > **Monitor Turning**.
The Monitor Tuning tab is displayed.
2. In the **Customer Persona** field, enter a customer persona.
3. (Optional) For help writing your persona, click **View Example**.
The **Example Customer Persona** window is displayed. Click **Grat, Thanks** to close the window.
4. In the **Regions Selection** section, select the regions to monitor.
5. In the **Category Pricing Range Inputs** **Min** and **Max** fields, enter the approximate minimum and maximum prices for items in each category.
6. Click **Save**.
Monitor Tuning is updated. The next monitoring cycle will reflect your tuning changes.
## Related Topics
* [Brand Config Overview](/product-agent/admin-settings/settings/brand-config-overview)
* [Brand Config Setup](/product-agent/admin-settings/settings/brand-config-setup)
# Taxonomy Overview
Source: https://developer.fabric.inc/product-agent/admin-settings/settings/taxonomy-overview
Learn how defining categories and attributes creates a structured catalog that supports consistent, high-quality product enrichment.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Taxonomy** menu in Settings is where you define the catalog structure and product data patterns that Product Agent uses as reference when enriching your products in Activate.
Taxonomy establishes two core elements:
* **Categories**, which define how products are grouped in your catalog
* **Attributes**, which define the types of product details associated with those products
Together, categories and attributes provide the structure Product Agent needs to interpret your catalog consistently and generate more accurate enriched content.
Setting up your taxonomy ensures that your product data is organized and ready for enrichment.
## Categories
**Categories** determine how your products are organized in Product Agent.
You can create a hierarchical category tree that reflects your merchandising structure, starting with top-level categories and adding subcategories as needed. There is no fixed limit to the number of hierarchy levels, allowing you to create simple or highly detailed product structures.
A well-defined category tree:
* Groups related products logically
* Supports consistent classification across your catalog
* Ensures every product has a clear placement
For example, a furniture company’s category hierarchy might look like the following:
* Furniture
* Tables
* Dining tables
* Oak dining table seating 4
* Oak dining table seating 6
* Coffee tables
* Oak coffee table
* Oak coffee table, glass top
* Chairs
* Dining chairs
* Oak dining chair
* Oak dining chair, with arms
* Office chairs
* Computer chair
* Computer chair, reclining
Each category can be created, renamed, nested beneath another category, or deleted. Your taxonomy should account for all products in your catalog so enrichment and structured data remain consistent.
## Attributes
**Attributes** are the details that describe the products in your catalog, such as the color, material, or dimensions. Defining attributes helps Product Agent understand what information exists in your catalog and how it should be interpreted during enrichment.
Each attribute represents the data type it stores. Some attributes contain text, others contain numbers or dates, and some are limited to a specific set of options. Certain attributes can store multiple values for a single product.
The following table displays the available attribute types, their descriptions, and their uses.
| **Data Type** | **Description** | **Example Use Case** |
| :----------------- | :---------------------------------------------- | :--------------------------------------- |
| **String** | Free-form text values. | Description, material, care instructions |
| **Number** | Numeric values. | Height, weight, price |
| **Date** | Calendar-based values. | Release date, expiration date |
| **Boolean** | True/false values. | Free shipping, eco-friendly |
| **List of values** | A predefined list allowing multiple selections. | Available colors, compatible devices |
| **Enum** | A predefined list allowing a single selection. | Size (small, medium, large) |
## Related Topics
* [Creating Categories](/product-agent/admin-settings/settings/creating-categories)
* [Creating Attributes](/product-agent/admin-settings/settings/creating-attributes)
* [Importing CSV Files Using API Endpoints](/product-agent/developer-guides/postman-import-csv-guide)
# Get attribute import status
Source: https://developer.fabric.inc/product-agent/api-reference/attributes/get-import-status
product-agent.openapi get /v2/attributes/import/{import_id}
Retrieve the current status and results of an attribute import created using `POST /v2/attributes/import`.
# Import attributes from CSV
Source: https://developer.fabric.inc/product-agent/api-reference/attributes/import-attributes
product-agent.openapi post /v2/attributes/import
Upload a CSV file to import attribute definitions for the specified brand/domain. The import is processed asynchronously.
**Expected CSV columns**
- **attribute_name** (required): Attribute display name
- **attribute_key** (optional): Machine-readable key
- **description** (optional): Attribute description
- **data_type** (optional): Data type
- **scope** (optional): Attribute scope
- **enum_values** (optional): Allowed values, pipe-delimited (e.g. `S|M|L`)
- **allow_ai_content** (optional): Allow AI to generate values (`true`/`false`)
- **source** (optional): `MERCHANT` or `GOLD_STANDARD`
- **guideline_reason** (optional): Guidance for AI enrichment
# Create Access Token
Source: https://developer.fabric.inc/product-agent/api-reference/authorization/get-token
product-agent.openapi post /platform/v1/auth/token
Generates an access token using the client credentials flow. Tokens expire every 60 minutes.
To authenticate, provide your `client_id`, `client_secret`, and set `grant_type` to `client_credentials`.
If you do not have a `client_id` and `client_secret`, contact fabric support to request API access credentials.
# Get category import status
Source: https://developer.fabric.inc/product-agent/api-reference/categories/get-import-status
product-agent.openapi get /v2/categories/import/{import_id}
Retrieve the current status and results of a category import created using `POST /api/v2/categories/import`.
# Bulk Import Categories
Source: https://developer.fabric.inc/product-agent/api-reference/categories/import-categories
product-agent.openapi post /v2/categories/import
Upload a CSV file to import category data for the specified brand/domain. The import is processed asynchronously.
**Expected CSV columns**
- **name** or **category_name** (required): Category name
- **parent_id** or **parent_category_id** (optional): Parent category ID
- **breadcrumb** (optional): Full breadcrumb path
- **inferred_category** (optional): AI-inferred category
# Bulk Import Products
Source: https://developer.fabric.inc/product-agent/api-reference/products/bulk-import-products
product-agent.openapi post /v2/taxonomy-workflow
Upload a CSV file of products to start a taxonomy workflow for the specified brand/domain.
**Expected CSV columns**
- **title** (required): Product title
- **breadcrumb** (required): Category breadcrumb path
- **sku** (optional): Product SKU
- **category** (optional): Product category
- **description** (optional): Product description
- **images** (optional): Image URLs
- **attribute.\*** (optional): Custom attributes (e.g. `attribute.color`)
# Run catalog enrichment
Source: https://developer.fabric.inc/product-agent/api-reference/products/enrich-products
product-agent.openapi post /v3/product-agent-api/catalog-workflow
Creates a catalog enrichment workflow for an uploaded product catalog. Use the `output_file_id` returned from the review and approve product import endpoint as the `catalog_file_id` in the request body.
# Get catalog enrichment status
Source: https://developer.fabric.inc/product-agent/api-reference/products/get-enrichment-status
product-agent.openapi get /v3/product-agent-api/catalog-workflow/{workflow_id}
Retrieves the current status of a catalog enrichment workflow.
# Get taxonomy workflow status
Source: https://developer.fabric.inc/product-agent/api-reference/products/get-import-status
product-agent.openapi get /v2/taxonomy-workflow/{workflow_id}
Retrieve the current status and progress of a taxonomy workflow created using `POST /v2/taxonomy-workflow`.
# Complete review and approve product import
Source: https://developer.fabric.inc/product-agent/api-reference/products/review-and-approve-import
product-agent.openapi post /v2/taxonomy-workflow/{workflow_id}/review/complete
Completes review for a taxonomy workflow and approves the uploaded product file for downstream enrichment.
# Product Import Developer Guide
Source: https://developer.fabric.inc/product-agent/developer-guides/postman-import-csv-guide
Import product data through CSV upload to start enrichment.
## Prerequisites
In order to upload products, you must already have created your category taxonomy. For information on creating categories in the Product Agent UI, see [Creating Categories](/product-agent/admin-settings/settings/creating-categories).
## Step 1: Postman Environment Setup
This part of the guide walks you through setting up a Postman environment to interact with the Product Agent API.
### Create a new environment
1. Open Postman.
2. Click **Environments** in the sidebar.
3. Click **+** to create a new environment.
Provide a name such as *Product Agent - Prod*.
### Add the environment variables
Add the following variables to your environment.
`access_token` and the various defined IDs will be empty initially — they will be populated after making API calls.
| Variable | Description | Example |
| --------------------- | ---------------------------------------------------------------- | :----------------------------------------------- |
| `baseUrl` | Base URL for the API | `https://commerceos.aiagents.fabric.inc/api` |
| `authUrl` | Auth URL for the API | `https://commerceos.aiagents.fabric.inc` |
| `access_token` | Token used for authenticated requests | `eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9...` |
| `taxonomyWorkflowId` | ID of the taxonomy workflow (used for status endpoints) | `a1462a91-f733-45fe-993b-5d0353f33ee3` |
| `attributeWorkflowId` | ID of the attribute workflow (used for status endpoints) | `a1462a91-f733-45fe-993b-5d0353f33ee3` |
| `clientId` | Client ID provided by fabric. Used to create an access token | `svc_peyFD9BXPRrZymhjJtFuYr7L3Ai` |
| `clientSecret` | Client secret provided by fabric. Used to create an access token | `cw_GWve_9oI_aJKd-xcE7uuZJpr-WqfnRpDPznGNVI-fOc` |
| `domain` | Brand domain or identifier for scoping requests | `acme.com` |
## Step 2: Authentication
This part of the guide walks you through authenticating with the Product Agent API and storing your access token in Postman for future requests.
### Create and set the access token
To authenticate, send a POST request to the token endpoint using your `clientId` and `clientSecret`. If you do not have these credentials, contact fabric support.
Tokens expire every 60 minutes.
```shell theme={null}
curl --location '{{authUrl}}/platform/v1/auth/token' \
--header 'Content-Type: application/json' \
--header 'Accept: application/json' \
--data '{
"grant_type": "client_credentials",
"client_id": "{{clientId}}",
"client_secret": "{{clientSecret}}"
}'
```
```json theme={null}
{
"access_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiJzdmNfcGV5SkQ5DlhQUnRadW1oakp0RnVZcjdMbEFvIiwidHlwZ56UIkjnZpY2UiLCJjb21wYW55X2lkIjoiNjhiNzI3NjhlYzc1NGVjNzA2MjhkNTM5Iiwic2NvcGVzIjpbXSwiYnJhbmRfaWRzIjpbImNvbnRhaW5lcnN0b3JlLmNvbSJdLCJqdGkiOiI5MGQ3YjQ0NC03MTQxLTRhMDUtYWM4MC03OGJjZTMxZGI5OGUiLCJpYXQiOjE3NzQ1NDM2NDEsImV4cCI6MTc3NDU0NzI0MX0.kc9RQSirU2vIVl5oRTmuEBoaF1tmAeYB3KH87KAHJpM",
"token_type": "bearer",
"expires_in": 3600,
"scope": ""
}
```
To automatically save the `access_token` for future requests:
1. Go to the **Scripts** tab in your Postman request.
2. Select **Post-response**.
3. Add the following script:
`pm.environment.set("access_token", pm.response.json().access_token);`
With this script, the `access_token` environment variable is automatically updated each time a request to the endpoint is made.
## Step 3 (Optional): Import Attribute Definitions
You can optionally import attribute definitions for enrichment and mapping workflows. For information on creating attributes in the Product Agent UI, see [Creating Attributes](/product-agent/admin-settings/settings/creating-attributes).
This endpoint accepts a CSV file and creates an import job you can monitor using the returned `import_id`. We will save this to our environment as `attributeWorkflowId`.
### Expected CSV columns for each attribute definition
| Column | Required | Description |
| :----------------- | -------: | :---------------------------------------------------- |
| `attribute_name` | Yes | Attribute display name |
| `attribute_key` | No | Machine-readable key |
| `description` | Yes | Attribute description |
| `data_type` | No | Attribute data type |
| `scope` | No | Attribute scope |
| `enum_values` | No | Allowed values, pipe-delimited, for example `S\|M\|L` |
| `allow_ai_content` | No | Allow AI to generate values (`true` or `false`) |
| `source` | No | `MERCHANT` or `GOLD_STANDARD` |
| `guideline_reason` | No | Guidance for AI enrichment |
### Request
```shell theme={null}
curl --location '{{baseUrl}}/v2/attributes/import' \
--header 'Authorization: {{access_token}}' \
--header 'domain: {{domain}}' \
--form 'file=@"/D:/Demo CSV files/attributes.csv"'
```
To automatically save the `attributeWorkflowId` for future requests:
1. Go to the **Scripts** tab in your Postman request.
2. Select **Post-response**.
3. Add the following script:
`pm.environment.set("attributeWorkflowId", pm.response.json().import_id);`
```json theme={null}
{
"import_id": "e2716438-b763-4be8-82d2-36abe0cb92b1",
"name": "attribute.csv",
"status": "PENDING",
"total_rows": 36,
"processed_rows": 0,
"created_count": 0,
"updated_count": 0,
"skipped_count": 0,
"failed_count": 0,
"input_filename": "attribute.csv",
"input_file_url": null,
"error_file_url": null,
"error_message": null,
"errors": [],
"started_at": null,
"completed_at": null,
"created_at": "2026-03-12T18:04:03.879876Z"
}
```
### Check the import status
After the upload succeeds, use the saved `attributeWorkflowId` to check workflow status with:
```shell theme={null}
curl --location '{{baseUrl}}/v2/attributes/import/{{attributeWorkflowId}}' \
--header 'Authorization: {{access_token}}' \
--header 'domain: {{domain}}'
```
```json theme={null}
{
"import_id": "e2716438-b763-4be8-82d2-36abe0cb92b1",
"name": "attribute.csv",
"status": "COMPLETED",
"total_rows": 36,
"processed_rows": 36,
"created_count": 35,
"updated_count": 0,
"skipped_count": 0,
"failed_count": 1,
"input_filename": "attribute_test.csv",
"input_file_url": "https://product-agent-data-prod-ue2.s3.amazonaws.com/attribute/vessel/e2716438-b763-4be8-82d2-36abe0cb92b1.csv...",
"error_file_url": "https://product-agent-data-prod-ue2.s3.amazonaws.com/attribute/vessel/e2716438-b763-4be8-82d2-36abe0cb92b1_errors.csv?X-...",
"error_message": null,
"errors": [
"Row 35: invalid data_type 'BOOLREAN'"
],
"started_at": "2026-03-12T18:04:03.998437Z",
"completed_at": "2026-03-12T18:04:04.724584Z",
"created_at": "2026-03-12T18:04:03.879876Z"
}
```
In this example, an error was flagged in the upload. You can resolve and publish updates to just that row or re-upload the entire file. Rows with errors are skipped.
Once completed, you can review the attribute definitions you uploaded in the UI.
1. Log in to Product Agent.
2. In the left nav, click **Settings**.
The **Settings** menu is displayed.
3. Click **Taxonomy**.
The **Categories** tab is displayed by default.
4. Click **Attributes**.
Here you can review your attribute definitions.
## Step 4: Import Your Products
After importing your categories and attribute definitions, you can upload products. During processing, products will be automatically mapped to your categories and aligned with the attribute definitions you’ve provided.
This endpoint accepts a CSV file and creates a taxonomy workflow you can monitor using the returned workflow `id`. We will save this to our environment as `taxonomyWorkflowId`.
### Expected CSV columns
| Column | Required | Description |
| :------------ | -------: | :----------------------------------------------- |
| `title` | Yes | Product title |
| `breadcrumb` | Yes | Category breadcrumb path |
| `sku` | No | Product SKU |
| `category` | No | Product category |
| `description` | No | Product description |
| `images` | No | Image URLs |
| `attribute.*` | No | Custom attributes, for example `attribute.color` |
### Request
```shell theme={null}
curl --location '{{baseUrl}}/v2/taxonomy-workflow?auto_enrich=true' \
--header 'Authorization: {{access_token}}' \
--header 'domain: {{domain}}' \
--form 'file=@"/D:/Demo CSV files/products.csv"'
```
To automatically save the `taxonomyWorkflowId` for future requests:
1. Go to the **Scripts** tab in your Postman request.
2. Select **Post-response**.
3. Add the following script:
`pm.environment.set("taxonomyWorkflowId", pm.response.json().workflow_id);`
```json theme={null}
{
"id": "a1462a91-f733-45fe-993b-5d0353f33ee3",
"brand_id": "268465dd-71f5-4179-959c-9bac01029451",
"status": "PENDING",
"workflow_type": "ATTRIBUTE_AND_CATEGORY_MAPPING",
"import_document_path": "taxonomy-workflow/vessel/d6441888-f25d-4acb-8898-07808526ecdb/d6441888-f25d-4acb-8898-07808526ecdb_input.csv",
"created_at": "2026-03-12T17:54:07.374339Z",
"updated_at": "2026-03-12T17:54:07.374339Z",
"total_category_mappings": 0,
"total_attribute_mappings": 0,
"workflow_id": "a1462a91-f733-45fe-993b-5d0353f33ee3",
"workflow_status": "PENDING",
"tenant_id": "268465dd-71f5-4179-959c-9bac01029451",
"workflow_phase": "PRE_PROCESS",
"domain": "acme.com",
"total_items": 0,
"processed_items": 0,
"successful_items": 0,
"failed_items": 0,
"total_chunks": 0,
"processed_chunks": 0,
"input_file_id": "69b2fdbfd1dcca3911c666e6",
"output_file_id": null,
"error_file_id": null,
"review_required": false,
"review_task_id": null,
"enrichment_workflow_id": null,
"error_message": null
}
```
### Check the import status
After the upload succeeds, use the saved `taxonomyWorkflowId` to check workflow status with:
```shell theme={null}
curl --location '{{baseUrl}}/v2/taxonomy-workflow/{{taxonomyWorkflowId}}' \
--header 'Authorization: {{access_token}}' \
--header 'domain: {{domain}}'
```
```json theme={null}
{
"id": "2345943b-91dd-4a3b-a923-b5c524049dcc",
"brand_id": "268465dd-71f5-4179-959c-9bac01029451",
"status": "COMPLETED",
"workflow_type": "ATTRIBUTE_AND_CATEGORY_MAPPING",
"import_document_path": "taxonomy-workflow/vessel/3c007843-d82b-4805-93f1-1afef9d1968e/3c007843-d82b-4805-93f1-1afef9d1968e_input.csv",
"created_at": "2026-03-18T20:35:03.936559Z",
"updated_at": "2026-03-18T20:35:16.658894Z",
"total_category_mappings": 4,
"total_attribute_mappings": 0,
"workflow_id": "2345943b-91dd-4a3b-a923-b5c524049dcc",
"workflow_status": "COMPLETED",
"tenant_id": "268465dd-71f5-4179-959c-9bac01029451",
"workflow_phase": "PRE_PROCESS",
"domain": "acme.com",
"total_items": 0,
"processed_items": 0,
"successful_items": 0,
"failed_items": 0,
"total_chunks": 0,
"processed_chunks": 0,
"input_file_id": "69bb0c78826cb2bb154da262",
"output_file_id": null,
"error_file_id": null,
"review_required": false,
"review_task_id": null,
"enrichment_workflow_id": null,
"error_message": null
}
```
Wait for the status to be set to `COMPLETED`.
Once processing is complete, your products will appear in Activate. Based on your category structure, products will be organized accordingly—for example, sweaters under *Tops > Sweaters* and t-shirts under *Tops > T-Shirts*.
# Adding a Competitor for Monitoring
Source: https://developer.fabric.inc/product-agent/monitor/add-a-competitor
Compare your categories to competitors in AI shopping to uncover enrichment and category strategy opportunities.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
Adding competitors helps you understand how your products are performing in AI-driven shopping experiences relative to others in your space. This context can highlight opportunities to improve enrichment, strengthen citations, or refine category strategy.
1. Log into **fabric NEON**.
The **Visibility** page for **Monitor** is displayed by default.
2. Click **Add Competitor**.
The **Add Competitor** window is displayed.
3. In the **Competitor Website URL** field, enter a competitors URL.
A maximum of 5 competitors can be added.
4. Click **Add**.
The competitor is added as a benchmark so you can compare your products against their visibility performance over time. Once added, competitor data may take about 1 minute to load, and up to 5 minutes for larger catalogs.
## Related Topics
* [Overview](/product-agent/monitor/overview)
* [Adding a Monitored Category](/product-agent/monitor/add-a-monitored-category)
* [Category Performance Report](/product-agent/monitor/category-performance-report)
* [Prompts](/product-agent/monitor/prompts)
* [Adding a Prompt](/product-agent/monitor/add-a-prompt)
* [Recommendations](/product-agent/monitor/recommendations)
# Adding a Monitored Category
Source: https://developer.fabric.inc/product-agent/monitor/add-a-monitored-category
Use the Monitor tab to add categories you want to track, helping you follow visibility trends and focus your enrichment efforts.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
By adding categories to track in the Visibility tab, you can follow visibility trends and pinpoint where enrichment efforts will be most effective.
## Procedure
1. Log into **fabric NEON**.
The **Visibility** page for **Monitor** is displayed by default.
2. Scroll down to the **Monitored Categories** section and click **Add Category to Monitor**.
The **Add Categories to Monitor** window is displayed.
3. Select one or more categories.
4. Click **Update**.
The selected categories are now tracked and displayed in the **Monitored Categories** section of the **Monitor** tab.
## Related Topics
* [Overview](/product-agent/monitor/overview)
* [Adding a Competitor for Monitoring](/product-agent/monitor/add-a-competitor)
* [Category Performance Report](/product-agent/monitor/category-performance-report)
* [Prompts](/product-agent/monitor/prompts)
* [Adding a Prompt](/product-agent/monitor/add-a-prompt)
* [Recommendations](/product-agent/monitor/recommendations)
# Adding a Prompt
Source: https://developer.fabric.inc/product-agent/monitor/add-a-prompt
Create and manage custom search prompts to measure product visibility in AI-powered shopping engines
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
Product Agent automatically generates a set of prompts based on your enriched product data. These prompts reflect real-world, natural-language search queries that customers might use in AI-powered shopping experiences.
While Product Agent can generate high-quality prompts on its own, you can also add custom prompts to track performance for specific queries that are important to your business or brand.
## Procedure
1. Log into **fabric NEON**.
The **Visibility** page for **Monitor** is displayed by default.
2. Click the **Prompts** tab.
The **Prompts** page is displayed.
3. Do one of the following:
* Enter a prompt using the **Add a new prompt to monitor** text field.
* Click the **Add Prompt** button.
4. Select a product category for the prompt.
5. Do one of the following:
* If you used the text field to add a prompt, hit enter.
* If you used the **Add Prompt** button, click **Add**.
The prompt is added to your **Monitored Search Prompts** table.
## Using the Prompt Suggestion UI
The prompt Suggestion UI helps you generate prompts. These prompts can be manually edited or added as-is.
1. On the **Prompts** page, click **Prompt Suggestions**.
The **Customize & Add Suggested Prompts** window is displayed.
2. In the **Select a Prompt Intent** section, use the **Category** field to select a category.
3. Do one or more of the following:
* **Price Range**: Generates prompts that include a specific price range.
* **Region**: Generates prompts that include a geographic region.
For example, if you select the Helmets category, set the Region to Canada, and choose a Price Range between 100 and 500, the generated prompts would resemble:
*“What are the best motorcycle helmets for daily commuting, priced between CAD 100 and CAD 500, available in Canada, and where can I purchase them?”*
4. Click **Generate Prompts**.
Product Agent automatically generates prompts based on your prompt intent criteria.
5. Optionally, you can edit any of the generated prompts by clicking the edit icon.
For unbiased visibility results, avoid using your own brand name in prompts.
6. Click the plus icon to add the prompt to your **Monitored Search Prompts**.
## Related Topics
* [Overview](/product-agent/monitor/overview)
* [Adding a Monitored Category](/product-agent/monitor/add-a-monitored-category)
* [Adding a Competitor for Monitoring](/product-agent/monitor/add-a-competitor)
* [Category Performance Report](/product-agent/monitor/category-performance-report)
* [Prompts](/product-agent/monitor/prompts)
* [Recommendations](/product-agent/monitor/recommendations)
# Category Performance Report
Source: https://developer.fabric.inc/product-agent/monitor/category-performance-report
View detailed category performance, including visibility, content depth, rankings, and AI search insights, with tools to compare against leaders or competitors and refine enrichment strategy.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Category Details** page provides an in-depth view into how a specific product category performs across AI-powered search and shopping platforms. Here, you can review high-level scores, explore visibility and content depth analysis, compare your category against market benchmarks or chosen competitors, and examine how common search prompts influence your ranking. This report reveals your category’s strengths and gaps across AI engines by comparing your performance with both fabric’s category leaders and your chosen competitors, providing a multi-dimensional view of your market position.
## Performance Overview
This section summarizes your category’s core performance metrics:
* **Visibility Score** (0–100) — Measures how easily your category is discovered across AI search platforms.
* **Content Depth** (0–100) — Evaluates the completeness and richness of product information within the category.
* **Category Ranking** — Indicates how your category ranks in AI search relative to others in the same space.
* **Top AI Channel** — Shows which AI engine (Gemini, ChatGPT, or Perplexity) your category performs best on.
These metrics provide a quick snapshot of your overall category health and where to focus improvement efforts.
## Analysis
The **Analysis** section includes two tabs, **Visibility** and **Content Depth**, each offering tailored insights into your category’s performance.
### Visibility tab
The Visibility tab (selected by default) allows you to examine how well your category surfaces in AI-driven shopping experiences.
You can configure the analysis by:
* Choosing a comparison group:
* Category Leaders
* Competitors
* Toggling aggregate scores:
* View consolidated vs. individual AI-engine results.
Visibility ranges from no visibility (blank) to low (red), medium (yellow), and high (green). All visibility metrics are presented in a table, organized by AI channel.
### Content depth tab
Switching to the Content Depth tab reveals how comprehensive and well-structured your category’s product content is. The Content depth metric displays pre-optimization and post-optimization prediction scores. You can click **Optimize Category** to start the **Product Enrichment** process in [Activate](/product-agent/activate/overview).
Category depth is calculated by sampling products within the category and evaluating them across key metrics. The table supports the same options as the Visibility tab:
* Choosing a comparison group:
* Category Leaders
* Competitors
Content Depth Metrics include:
| Row | Description |
| :--------------------- | :------------------------------------------------------------------------------ |
| **Attribute Coverage** | Measures how many key attributes your product listings include. |
| **FAQ Coverage** | Assesses the presence and completeness of question-and-answer content. |
| **Keyword Coverage** | Evaluates how effectively category-relevant keywords appear in product content. |
| **Citations** | Indicates whether claims or details are backed by trusted sources. |
| **Image Alt Tags** | Measures the presence and usefulness of alt text for product images. |
These insights highlight gaps in content quality and guide improvements that enhance category performance across AI systems.
## Tracked Search Prompts
The **Tracked Search Prompts** section displays the prompts used to evaluate your search visibility. The table includes:
**Prompt** — The exact phrase used during analysis.
**AI Engines** (Perplexity, Gemini, ChatGPT) — Each column shows visibility scores using the same color indicators as other tables.
You can toggle the table between:
**Visibility Scores** — Numeric evaluations of how well your category appears.
**Ranking View** — Displays the top five businesses surfaced for each prompt.
### Adding a New Prompt
fabric automatically generates five baseline prompts for each category.
For unbiased visibility results, avoid using your own brand name in custom prompts.
1. Click **+ Add Prompt**
The **Add Prompt** window is displayed.
2. In the **Prompt** field, enter a prompt.
The **Category** field is preselected since you are adding a prompt within that category’s detailed view.
3. Click **Add**
Your prompt is added to the **Tracked Search Prompts** table for ongoing visibility monitoring.
## Related Topics
* [Overview](/product-agent/monitor/overview)
* [Adding a Monitored Category](/product-agent/monitor/add-a-monitored-category)
* [Adding a Competitor for Monitoring](/product-agent/monitor/add-a-competitor)
* [Prompts](/product-agent/monitor/prompts)
* [Adding a Prompt](/product-agent/monitor/add-a-prompt)
* [Recommendations](/product-agent/monitor/recommendations)
# Experiment Performance
Source: https://developer.fabric.inc/product-agent/monitor/experiment-performance
Analyze experiment results and compare performance over time.
After you create and run an Experiment in [Activate](/product-agent/activate/activating-products/experiments), you can view its performance in Monitor.
The **Performance** tab in Monitor shows how enriched products (variant) perform compared to their original versions (control) over time. This helps you understand the impact of enrichment and decide how to apply changes across your catalog.
## Experiment Analytics
Each experiment includes a set of graphs that compare variant and control performance over time.
* The blue line represents the enriched products (variant)
* The orange line represents the original products (control)
* The vertical line marks the treatment date, when enriched products were published
* The horizontal axis shows days from treatment
Each experiment includes the following graphs:
| **Graph** | **Vertical axis** | **Use** |
| :------------------------- | :---------------------------------------------------------------------------------------------------------- | :----------------------------------------------------------------------- |
| **Total Sessions** | **Effect (%)**, which reflects the percentage change in sessions compared to the control group | Evaluate how enrichment impacts product visibility and traffic over time |
| **Total Page Views** | **Effect (%)**, which reflects the percentage change in page views compared to the control group | Evaluate how enrichment affects product engagement and discovery |
| **Add to Cart Rate Trend** | **Add to Cart Rate (%)**, which reflects the percentage of product views that result in add-to-cart actions | Evaluate how enrichment impacts conversion behavior |
### Products in the experiment
The **Products** tab in the **Performance** view displays all products included in the experiment.
This is a read-only view that allows you to review which products are included in the experiment and verify the scope of your results.
## Prerequisites
* Ensure you have created at least one [Experiment](/product-agent/activate/activating-products/experiments) in Activate.
## Procedure
The following procedure details how to access data in **Performance**.
### Viewing experiment metrics
1. In the left menu, click **Product Agent** > **Monitor**.
The **Monitor** page is displayed on the **Visibility** menu.
2. Click **Performance**.
The **Performance** menu is displayed.
It shows all experiments, including their status, start and end dates, and progress.
3. To view the details of an experiment, hover over its title and click **View Details**.
The experiment is displayed on the **Analytics** menu.
4. (Optional) To view the products in the experiment:
* Click the button at the top-right to download a CSV file of the products in the experiment.
* Click **Products** to view the products in the experiment, broken down by category.
## Related Topics
* [Product Activation](/product-agent/activate/activating-products/introduction)
* [Experiments](/product-agent/activate/activating-products/experiments)
* [Importing Products](/product-agent/activate/activating-products/importing-products)
# Visibility
Source: https://developer.fabric.inc/product-agent/monitor/overview
Monitor category search visibility across multiple AI channels
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Visibility** page displays results for how your monitored category's perform in generative search engines and shopping assistants. It works directly with your current product catalog (no changes required) so you can quickly uncover trends, benchmarks, and opportunities.
You can use **Visibility** to:
* See how your AI visibility is changing over time
* Compare performance across monitored product categories
* Track how your catalog performs against competitors and category leaders
* View monitored categories, including detailed insights into content depth and tracked prompts
## Visibility
The Visibility section shows your aggregate visibility score across monitored categories and AI search engines. By default, the **Channel** tab highlights how you compare to the category leaders (aggregate), and you can optionally add up to five competitors to benchmark your performance.
Visibility ranges from no visibility (blank) to low (red), medium (yellow), and high (green). All visibility metrics are presented in a table, organized by AI channel.
The **Trend** view visualizes changes in overall visibility over time. When selected, it switches from the summary view to a graph that shows visibility score trends for category leaders or selected competitors, depending on your comparison settings. This view reflects overall visibility across monitored categories and AI search engines, including Perplexity, ChatGPT, and Gemini.
## Monitored Categories
You can use the **Visibility** tab to track how your product categories are performing across AI-powered search and shopping platforms. By selecting specific categories to monitor, you can follow changes in visibility over time, identify top-performing AI channels for each category, and benchmark your performance against competitors.
Category-level tracking helps you focus your enrichment strategy where it matters most, whether that's improving underperforming categories or reinforcing visibility in high-performing ones.
Categories appear in a table that shows:
| Column | Description |
| :------------------- | :-------------------------------------------------------------------------------- |
| **Rank** | Shows how this category ranks relative to the rest of your categories. |
| **Category** | The product category being monitored, such as *running shoes* or *kitchen tools*. |
| **Top AI Channel** | The generative search engine where your products in this category perform best. |
| **Visibility Score** | The current visibility score for the category. |
Hovering over a monitored category allows you click **View** to see more detailed insights, including tracked prompts, citation sources, and competitor benchmarks for that category.
## Related Topics
* [Adding a Monitored Category](/product-agent/monitor/add-a-monitored-category)
* [Adding a Competitor for Monitoring](/product-agent/monitor/add-a-competitor)
* [Category Performance Report](/product-agent/monitor/category-performance-report)
* [Prompts](/product-agent/monitor/prompts)
* [Adding a Prompt](/product-agent/monitor/add-a-prompt)
* [Recommendations](/product-agent/monitor/recommendations)
# Prompts
Source: https://developer.fabric.inc/product-agent/monitor/prompts
Understand how your products surface in AI-powered shopping queries and monitor prompt-level visibility and competitiveness.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Prompts** page in **Monitor** shows how your product categories perform in response to search prompts used in AI-powered shopping experiences. These prompts simulate real-world queries, such as “lightest running shoes” or “most durable kitchen blender,” and help you understand where and how your categories are surfacing in generative search results. It also allows you to see competitive rankings and identify prompt-level coverage gaps you can address through enrichment.
Product Agent automatically generates five prompts based on your product data. They are designed to reflect the kinds of natural-language queries customers might enter into AI engines when shopping for categories like yours. Prompts are inferred from product titles, attributes, descriptions, and categories. In addition to the auto-generated set, you can also create your own prompts to track how specific phrases or use cases perform over time.
You can use Prompts to:
* View how well your categories appear for high-intent, natural-language prompts
* See how your rankings compare to competitors at the prompt level
## Viewing Data
The Prompts tab has controls to create prompts as well as sort them based on title, category, and performance in individual AI engines.
The Prompts tab includes several controls that allow you to adjust how prompts are grouped and displayed. You can group prompts by product category, or choose whether to view them by visibility score or competitive ranking. Additionally, you can use the search field to find a specific prompt.
### Filtering by a category
Use the **Categories** dropdown in the top-left corner to select one or more categories. This filters the monitored search prompts to show only prompts associated with the selected categories. By default, prompts from all categories are displayed.
### Prompt table
To the top-right of the prompt table is a **view by** toggle for **Visibility** and **Ranking**. This toggle changes what's displayed in the table below.
* **View by Visibility**: Displays color-coded indicators showing how well each AI engine surfaces your categories for the selected prompt.
* **View by Ranking**: Shows your search visibility ranking for each prompt in the selected category, along with the top five performers.
| Field | Description |
| :-------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Prompt** | The prompt used to monitor your category-level search visibility. |
| **Category** | Displays a prompt's associated product category, such as *running shoes* or *kitchen tools*. |
| **Perplexity (icon)** | Shows a measured value of your visibility on Perplexity: green for high visibility, yellow for medium, red for low, and no color when not visible. |
| **ChatGPT (icon)** | Shows a measured value of your visibility on ChatGPT: green for high visibility, yellow for medium, red for low, and no color when not visible. |
| **Gemini (icon)** | Shows a measured value of your visibility on Gemini: green for high visibility, yellow for medium, red for low, and no color when not visible. |
## Exploring Prompts
The **Prompt Explorer** lets you dive deeper into how individual prompts perform across AI search engines. Hover over a prompt and click **Explore Prompt** to open the **Prompt Explorer** window.
### Search sources view
By default, the **Prompt Explorer** opens in the **Search Sources** view and provides a breakdown of how AI search engines responded to a prompt, including:
* How your site ranked compared to competitors and other surfaced sites
* How many references were made for each site
### Search results view
Use the **Search Results** tab to understand product-level visibility:
* View which products were surfaced for the prompt
* Access direct links to surfaced products
* See results by AI search engine
Use the dropdown in the top-left corner to switch between:
* ChatGPT
* Gemini
* Perplexity
## Related Topics
* [Overview](/product-agent/monitor/overview)
* [Adding a Monitored Category](/product-agent/monitor/add-a-monitored-category)
* [Adding a Competitor for Monitoring](/product-agent/monitor/add-a-competitor)
* [Category Performance Report](/product-agent/monitor/category-performance-report)
* [Adding a Prompt](/product-agent/monitor/add-a-prompt)
* [Recommendations](/product-agent/monitor/recommendations)
# Recommendations
Source: https://developer.fabric.inc/product-agent/monitor/recommendations
View and compare category-level content depth metrics against competitors and leaders, with pre- and post-optimization performance insights.
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
The **Recommendations** page displays how comprehensive and well-structured your category’s product content is. The Content depth metric displays pre-optimization and post-optimization prediction scores. You can click **Optimize All Products** to start the **Product Enrichment** process in [Activate](/product-agent/activate/overview).
## Filtering Categories
Use the **Category** dropdown in the top-left corner to select a specific category. This filters the **Content Depth** metrics to show only metrics associated with the selected category.
By default, the aggregate of all categories is displayed.
## Content Depth
Content depth is calculated by sampling products within a category and evaluating them across key quality metrics. You can compare your performance against **Category Leaders** or selected **Competitors**. The content depth table highlights gaps in content quality and helps identify targeted improvements to strengthen category performance across AI search.
Content Depth Metrics include:
| Row | Description |
| :--------------------- | :------------------------------------------------------------------------------ |
| **Attribute Coverage** | Measures how many key attributes your product listings include. |
| **FAQ Coverage** | Assesses the presence and completeness of question-and-answer (FAQ) content. |
| **Keyword Coverage** | Evaluates how effectively category-relevant keywords appear in product content. |
| **Citations** | Indicates whether claims or details are backed by trusted sources. |
| **Image Alt Tags** | Measures the presence and usefulness of alt text for product images. |
## Related Topics
* [Overview](/product-agent/monitor/overview)
* [Adding a Monitored Category](/product-agent/monitor/add-a-monitored-category)
* [Adding a Competitor for Monitoring](/product-agent/monitor/add-a-competitor)
* [Category Performance Report](/product-agent/monitor/category-performance-report)
* [Prompts](/product-agent/monitor/prompts)
* [Recommendations](/product-agent/monitor/recommendations)
# Overview
Source: https://developer.fabric.inc/product-agent/overview
Product Agent is currently in beta, and we're [seeking partners](https://fabric.inc/contact-us) to help shape its future.
AI searches are changing how consumers find and evaluate products. Agents such as ChatGPT, Perplexity, and Gemini prioritize semantic understanding, contextual cues, and content richness. To compete in these environments, brands need to modernize how their product data is structured.
**Product Agent** helps you compete by improving content depth and enriching product information to make it accessible to AI agents and search platforms through two core workflows:
* [Monitor](/product-agent/monitor/overview): Measure and analyze how products appear in AI search contexts.
* [Activate](/product-agent/activate/overview): Enrich and deploy product data optimized for agentic commerce channels.
## Monitor: Track and Benchmark AI Visibility
**Monitor** helps you understand how your catalog performs in AI search results. It focuses on how well product content ranks and competes in real-world AI search prompts.
You can track catalog performance with fabric's built-in prompts, or your own custom prompts, and then benchmark against top-performing competitors in your category.
In addition to visibility tracking, Monitor provides content guidance: actionable recommendations to improve data completeness, structure, and relevance. These insights help prioritize content upgrades that will deliver the most visibility lift.
### Example use cases
* **Evaluate AI search readiness**: Understand how well your catalog is structured for AI search.
* **Category-level benchmarking**: Compare your brand’s visibility against competitors.
* **Prompt-based auditing**: Run custom prompts to evaluate how your listings perform in realistic searches.
## Activate: Enrich and Deploy Optimized Product Data
Once visibility gaps are identified, **Activate** helps you optimize product data and publish it to your sales channels. Activate provides tools to enrich your catalog using fabric's golden standard data set.
Features such as **taxonomy mapping** and **brand configuration** ensure that product data is not only enriched, but also consistent with your brand standards, and validation options allow you to retain editorial control.
### Example Use Cases
* **Onboard supplier catalogs at scale**: Map product data and fill in missing context automatically.
* **Enrich SKUs before launch**: Before pushing new products, activate richer metadata to increase visibility in AI searches.
## Workflow
Product Agent is designed for iterative use. A typical workflow looks like this:
1. **Monitor** product visibility across AI search prompts and engines.
2. Identify gaps in performance or content structure.
3. **Activate** enriched data to improve product relevance and visibility.
4. Re-monitor and refine based on updated results.
This continuous improvement ensures that your catalog evolves with AI search.