Data Sources

Data sources are the product data files you import into a project. They provide the raw information — product names, descriptions, prices, images, and more — that the AI uses to generate your content. Every project starts with at least one data source, and you can add as many as you need.

Supported Formats

XC AI Content Automation accepts two file formats for data sources:

• CSV (Comma-Separated Values) — A plain-text spreadsheet format. This is the most common way to import product catalogs and feeds.
• XML (Extensible Markup Language) — A structured data format often used by product feed providers and e-commerce platforms.

Both formats are parsed automatically on upload. In most cases, you do not need to configure a schema or mapping before importing -- the platform detects your columns automatically.

CSV Format

When you upload a CSV file, the platform uses a multi-strategy auto-detection engine to identify which columns in your file correspond to product fields. The following product fields are recognized:

Field	Description
Product ID / SKU	A unique identifier for the product
Product Name	The display name of the product
Short Description	A brief summary of the product
Long Description	A detailed product description (or body HTML)
Price	The product price
Images	One or more image URLs, typically separated by commas or pipes
Categories	The product category or category path

Platform-specific fields are also detected. For example, Shopify files may include Vendor, Product Type, and Tags columns, while Magento 2 files may use sku, description, and short_description as headers.

You do not need to include every column. The platform will work with whatever data you provide, but richer source data generally leads to better generated content.

Automatic Column Detection

The detection engine uses four strategies, applied in order of decreasing confidence:

1. Exact match -- Your column header matches the expected header for your platform exactly (e.g., "SKU" for WooCommerce, "Handle" for Shopify). This gives the highest confidence.
2. Alias match -- Your header matches a known alias in any supported language (e.g., "Artikelnummer" for product ID in German, "Nom" for product name in French). Aliases are recognized for English, German, French, Spanish, Portuguese, Italian, Slavic languages (Croatian, Serbian, Bosnian, Slovenian, Czech, Slovak, Polish), Nordic languages (Swedish, Norwegian, Danish, Finnish), and Dutch.
3. Keyword match -- Your header contains a keyword associated with a field (e.g., a header containing "description" maps to the description field).
4. Content analysis -- When headers alone are not enough, the platform examines sample data values. URLs suggest image fields, decimal numbers suggest price fields, long text or HTML suggests description fields.

When Detection Is Confident

If the engine confidently identifies both the Product ID and Product Name columns (confidence 90% or higher), the file is imported immediately using the detected mapping. No manual intervention is needed.

Column Mapping Dialog

If the engine is uncertain about any required mapping, a Column Mapping dialog appears before the import proceeds. This dialog shows:

• A table of detected fields with dropdown menus to select which CSV column maps to each field.
• Sample values from your file next to each mapping so you can verify the selection.
• Low confidence badges on any mapping the engine was less sure about, highlighted in amber so you can focus your attention there.
• Required fields (Product ID and Product Name) marked with an asterisk. The Import button is disabled until both are mapped.

Review the mappings, adjust any that look incorrect using the dropdown menus, and click Import to proceed. You can also set any field to Skip if you do not want to import it.

Uploading a File

To add a data source to your project:

1. Open your project from the dashboard.
2. In the Data Sources panel, click Add Source.
3. Select your CSV or XML file from your computer.
4. If column detection is confident, the import starts immediately. If not, the Column Mapping dialog appears for you to review and confirm the mappings.
5. Once confirmed, items are processed and imported automatically.

You will see a progress indicator while the file is being parsed. When processing is complete, the source appears in the Data Sources panel with a summary of the imported items.

Manual Entry

For small catalogs or one-off additions, you can add individual product entries manually instead of uploading a file. This is useful when you need to generate content for just a handful of products or want to test the platform before preparing a full data export.

To add a manual entry, open your project and use the manual entry option within the Data Sources panel. Fill in the product fields and save.

Viewing Sources

Each data source listed in your project displays the following information:

• File name — The name of the uploaded file (or "Manual Entry" for hand-entered items).
• Number of items — The total count of product records imported from that source.
• Processing status — The current state of the import, such as "Processing," "Completed," or "Error."

Click on any source to view its contents and inspect the individual product records that were imported.

Exporting Sources

You can export source data at any time for your own reference or for use outside the platform. This gives you a downloadable copy of the product records as they were imported, which can be helpful for auditing or comparison purposes.

To export, select a source from the Data Sources panel and choose the export option.

Deleting a Source

If a data source is no longer needed, you can remove it from your project. Deleting a source removes the imported product records associated with that file.

Importantly, deleting a source does not delete any content that has already been generated from it. Your generated product descriptions, titles, and other output remain intact.

To delete a source, select it in the Data Sources panel and confirm the deletion when prompted.

Tips

• Clean your data before upload. Remove duplicate entries, fix any character encoding issues, and verify that column headers are present. Clean input leads to better output.
• Use one source per product catalog or feed. This keeps your data organized and makes it easier to update or replace a specific feed later.
• Large files may take a moment to process. File parsing and item extraction happen in the background. You can continue working in your project while processing completes.
• You can upload multiple sources to the same project. This is useful when your product data comes from different systems or when you want to combine catalogs from multiple suppliers.