Introduction to Structured Attribute Enrichment

This guide walks you through importing and enriching your product database step by step. You'll learn how to prepare your import file, select enrichment options, and sync updated data back to your system.

In this article

1. Preparing import file
2. Import product data
3. Selecting products to enrich
4. Selecting fields to enrich and enrichment method
5. Review enriched attributes
6. Creating product descriptions
7. Syncing back to the database

Prepare Import File

Hypotenuse AI's Product Database requires a list of products to start. You can import products in multiple ways, including through an integration (if set up) or by uploading a file. This section covers file-based imports.

The database accepts CSV, XLSX, and XLS formats. To get started, create a file with the following headers:

• Unique ID: Ensure each product corresponds to a unique identifier, such as a SKU or internal identification value.
• Product title: A clear product name would assist with legibility, filtering, and AI enrichment.
• Brand (optional): Useful for identifying the product.
• Barcode (optional): UPC or EAN for enhanced product identification and enrichment.
• Model/part number (optional): Useful for differentiating products with similar names.

Filling in these product attribute fields will better help our AI product match during the Import mapping and mapping fields for product description stage.

Importing Product Data

To begin using the product database, click Import in the top left corner and select Structured CSV. Then, select the CSV file you would like to import.

Once the file is uploaded, click Continue to access the Import mapping field. This process maps your CSV file's fields to those in Hypotenuse AI's Product Database. There are two types of fields:

• Platform fields: Standard fields like Unique ID, Product title, Brand, etc. Assign the corresponding columns from your CSV to these fields.

• Custom fields: Where product attributes are stored, such as Length, Weight, Color, Temperature, Dimensions etc. Create a custom field by selecting the Add product attribute field and assigning the relevant datatype. (Read more about datatypes here)

After mapping all necessary fields, click Continue. Your product database will now be populated with data from your import file.

Selecting Products for Enrichment

To fill in missing attributes, use Hypotenuse AI’s Data Enrichment feature. This AI-powered tool populates missing values based on your taxonomy and data standardization format.

The uploaded CSV contains only basic product details (e.g., brand, product title, and image links).

In the below example, the attributes Color, Length, and Shape are currently empty.

To begin the enrichment process:

• Select individual products by ticking the checkbox on the left or click Select everything to enrich all products in the database.
• Click Enrich attributes.

Selecting Fields & Enrichment Method

Field selection

Select the fields you want to enrich and choose the enrichment method.

Enrichment types

• Web enrichment: AI scrapes the web for product attributes.
• Image enrichment: Analyzes images in the product database and extracts relevant details.
• Barcode enrichment: Uses barcode data to populate fields based on available information.
• PDF enrichment: Scrapes a PDF and fills out attributes with relevant elements.

Web enrichment

Suited for product attributes that can be easily located via internet sources (Eg. length, size, weight, etc.)

Once the relevant fields and enrichment types are selected, click Start enrichment.

Image enrichment

Best suited for visual attributes (e.g. color, pattern, etc.).

Following the previous steps, select Image enrichment and click Start enrichment.

PDF Enrichment

After selecting the fields you want to enrich, choose the PDF enrichment type and the corresponding Relevant unique identifiers.

Upload a PDF and then select Start enrichment.

Reviewing Enriched Attributes

The new attributes will be added to the database after enrichment. Before finalizing, review the enriched values to ensure accuracy.

Hypotenuse AI simplifies this step by generating a confidence score for each enriched value. Attributes requiring review are highlighted in yellow.

To resolve:

• Hover over the product field.
• A dropdown will appear. If the field is correct, click Resolve.

With this final step, your products are now enriched!

Generating Product Descriptions

Hypotenuse AI’s Product Catalog enables the seamless generation of product descriptions using enriched attributes.

To generate descriptions:

1. Select the enriched products.
2. Click Product descriptions under Settings.

3. In the Add new catalog window, enter a title and click Confirm.

4. Verify that the attributes are mapped correctly in the Map fields for product description window. This setup is needed only once; Hypotenuse AI will remember the mapping for future use.
5. Click Export to catalog.

6. In the Catalog tool, to generate product descriptions:
   • Click Bulk Generate to create descriptions for all products.
   • Or click Generate for a single item.

You should now see new product descriptions based on the enriched attributes.

Different brands may have specific product description guidelines and brand voices. Hypotenuse AI’s platform offers various AI writers who can tailor descriptions accordingly.

This setting can be managed in the Writers tab underneath Bulk actions.

Syncing Back to Database

After generating product descriptions, sync them back to the Product Database.

1. Click Export in the top left of the catalog and select Product Database.

2. Select the product(s) to export and press Next.

3. In the Map product database fields window, select the column where the product description should appear. Then click Export and Publish.

Upon returning to the product database, the newly generated descriptions should be visible.

With this process, managing your product database is seamless!

For further assistance, reach out to support@hypotenuse.ai—our team is happy to help!