E-commerce and retail · Sample scenario
Preparing content for the product catalog
Attributes are extracted from the supplier file, a description and SEO text are drafted, and after editor approval they are written to the catalog.
Problem
Writing descriptions and attributes for hundreds of new products by hand takes weeks.
How it works today
A content manager opens the supplier's Excel file, writes each description from scratch and copies attributes into the CMS by hand.
What this scenario changes
Attributes are extracted from the file, missing facts (size, material) are checked, a description and SEO text based only on available facts are drafted, and they are written to the CMS after editor approval.
The process step by step
- Supplier file: A new file is uploaded
- Attributes extracted: Structured fields from text
- Required fields checked: Size, material, color
- Are the facts complete?: All required fields filled
- Description and SEO text: Using available facts only
- Editor approval: Check before publishing
- Written to the catalog: Ready to publish in the CMS
Exception path
- Data requested from the supplier: The missing field is shown
- SKU on hold: Not published until the data arrives
The role of AI
Extracts attributes from unstructured supplier text and writes a description based only on available facts.
Human approval and exceptions
An editor approves every description before publishing; if a fact such as size or material is missing, AI does not invent it and the data is requested.
Frequently asked questions
Can AI make up facts about a product?
The description is written only from facts in the file; if a required fact is missing, the SKU is not published and the data is requested.
Are the texts published straight to the website?
No. Each text is written to the CMS after editor approval.
Which file formats are supported?
Usually Excel, CSV or Google Sheets; the exact format is agreed during the initial audit.