What is Web-based attribute enrichment?

Summary

Web-based attribute enrichment is an AI Configuration type in the PIM that fills Attribute values using information the AI finds on the live web. Other AI configurations generate content from data already in your catalog or translate existing values. Web-based attribute enrichment goes a step further: it searches public web sources for details about a product and proposes values for the Attributes you choose.

In short: given a product you already have, it finds reliable details about that product from public web sources and suggests them as Attribute values.

 

When to use it

Use Web-based attribute enrichment as a last line of defence, after you have used the other enrichment options available to you (shared catalog data, Assets, supplier documents, and standard AI generation). It is most valuable when the information you need is not already anywhere in your PIM but is publicly available online.

 

Expect it to be slower than standard generation

Because the AI performs a live web search, each Attribute takes roughly 10 to 30 seconds, compared with roughly 3 to 5 seconds for standard generation. It also costs more per call. Keep this in mind for high-volume bulk enrichment.

 

 

What it can fill

Web-based attribute enrichment supports a wide range of Attribute types:

  • Text and Text Area (for example a short specification or a product description)
  • Number, Measurement, Date, Yes/No, and Price
  • Simple select, Multi select, Reference Entity link, and Asset collection

For each Attribute type, the AI adapts how it works. Free-text Attributes are written as an original synthesis of what the AI found. Factual Attributes (such as a weight or a date) return the precise value. Option-based Attributes are matched to the options that already exist in your catalog.

 

How it works at a glance

An administrator creates a Web-based attribute enrichment configuration and chooses the target Attribute, the product context to use, and an instruction prompt.

  1. A user or an automation triggers the configuration on a product.
  2. The AI searches the web, then proposes a value together with the source websites it used.
  3. You review the suggestion, check the sources, and accept or reject it.
Proposed value on a product with an expandable list of source websites and accept and reject controls

 

Staying safe and compliant

Every suggestion is backed by the source websites the AI used, and several safeguards run automatically before a value is proposed. These include a publisher permission check, a customer-managed list of blocked websites, and a requirement that every value carries at least one source. To understand these protections in detail, see the related article on how Web-based attribute enrichment keeps your data safe and compliant.

 

Suggested use cases

Web-based attribute enrichment is most useful when the information you need is not already in your PIM but is publicly available online. A few common examples:

  • Filling technical specifications such as weight, dimensions, material, or power requirements, sourced from a manufacturer or distributor page.
  • Adding brand or manufacturer details when you have a model number or product title but the brand field is empty.
  • Drafting product descriptions for items where you have only basic data, letting the AI write an original summary from public sources for a reviewer to refine.
  • Enriching newly onboarded or supplier-provided products that arrive with sparse data, as a quick first pass before manual enrichment.
  • Completing factual Attributes like country of origin, release date, or a price reference, where a precise, source-backed value matters.
  • Matching option-based Attributes such as a category, feature tag, or Reference Entity link, where the answer must map to options that already exist in your catalog.

Remember to treat it as a last line of defence: use your existing data, Assets, and supplier documents first, and turn to Web-based attribute enrichment for the gaps that remain. Because every value is backed by sources and reviewed before saving, it works best on Attributes where you can quickly verify the suggestion.

Availability

Web-based attribute enrichment is enabled and available on all plans that have access to AI Configurations.