Skip to main content
Extracting product attributes with AI in WISEPIM Extracting product attributes with AI in WISEPIM Attributes are the structured specs that power filters, comparison tables, feeds, and faceted search, material, dimensions, color, weight, compatibility, and more. The problem is they are usually buried inside description text or missing entirely. Attribute enrichment uses AI to read each product’s existing content and extract those specs into clean, structured attribute fields. What it enables: make your catalog filterable and feed-ready. Structured attributes drive on-site faceted search, marketplace requirements, and the comparison data shoppers use to choose, all extracted automatically from text you already have.
Attribute enrichment costs 1 credit per product. Values are written to the attributes defined for that product’s family. A run that fills nothing is reported as no change, and its credit is refunded.

​
When to use it

  • Specs live in description text but not in structured fields, so shoppers can’t filter on them.
  • A marketplace or feed requires specific attributes (Google Shopping needs GTIN, color, size, material; Amazon has category-specific required fields).
  • Filters return incomplete results because many products are missing the attribute being filtered on.
  • As part of Auto-Fill, which extracts missing attributes using your default attribute prompt.

​
How to run it

1

Select products

Check products individually, or filter (for example, by a missing key attribute) and select all matching results.
2

Open the Enrich modal

Click Enrich with AI in the toolbar.
3

Choose Attributes

Pick Attributes as the enrichment type.
4

Select a prompt

Choose an attribute prompt from your Prompt Library. The prompt guides how values are found and formatted, and can name the exact attributes to fill (see Target specific attributes). Your default comes from the Attributes row in Enrich Prompts.No prompt of your own yet? Open the Curated Prompts tab. These are written by WISEPIM and ranked for your industry, so the best fit is near the top. Use Set as default on the one you like and it is pre-selected for every attribute run from then on.
5

Start Enrichment

Click Start Enrichment and track it in the Process Tracker. If a run stops early, use Resume there to finish only the products that were not processed yet, at no extra credit cost.

​
How extraction works

The AI reads each product’s title, description, attributes and its latest web research, then fills the attributes defined on the product’s family that have AI suggestions switched on. When the original description is empty, it reads the Enriched Description instead. Every value is checked before it is saved:
  • It needs a source. Each value must be quoted from the product data or the research. If the AI cannot point to it, the attribute stays empty. A blank is better than a guess.
  • It must fit the attribute. A value of the wrong type, outside the attribute’s options, or breaking its validation rules is rejected.
  • Your values stay. “Unknown” never clears a filled attribute, a missing yes/no never becomes “No”, and values a person entered are kept.
Big families are handled in parts, so every attribute gets a turn however many your family has. If one part fails, it is retried in two smaller halves, and the task log names any attribute that still could not be filled.
Attribute extraction is only as good as the source text. If descriptions are thin, run Web Research first, then extract attributes from the richer data.

​
Target specific attributes

An attribute prompt can name the attributes it writes. In the prompt form in the Prompt Library, pick them in Attributes to fill, for example only the color and material, or one FAQ text attribute. Leave it empty to fill every attribute that has AI suggestions switched on.

​
Choice attributes and new options

A single or multiple choice attribute can only take one of its options. When the AI finds a value that no option fits: Switch Let AI add new options on in the attribute’s settings in Managing Product Attributes. It works for attributes with their own option list, not for shared option sets. A choice attribute with no options and this switch off can never be filled by AI.

​
Why a field stayed empty

After you run Enrich with AI from a product’s Attributes tab, a notice lists the fields that stayed empty, for example “3 fields stayed empty”, with the reason for each: When nothing was saved at all, you see “AI enrichment didn’t change anything” with “No attributes found that the sources back up.”

​
Research can create attributes

Web research often finds specs your data model has no attribute for. In the product’s Research tab, Specs without an attribute lists them with a suggested name, type and the researched value. Tick the ones to keep and click Create and fill N attributes. See Web Research.

​
What good looks like

  • Consistent value formatting. “Stainless Steel” everywhere, not a mix of “stainless”, “S/S”, and “inox”. Ask the prompt to normalize.
  • Units included and consistent. Dimensions and weights carry their unit, in the same system across the catalog.
  • No invented values. A blank is better than a guess. Tell the prompt to leave an attribute empty if the value isn’t present in the source.
  • High coverage on filterable attributes. The attributes shoppers filter on (size, color, material) should be populated on nearly every product.

​
Act on what you find

Open the product’s Attributes tab and read the reason. Most often the attribute has no options, or the value found is not one of them. Add the options, or switch on Let AI add new options. Outcome: choice attributes fill on the next run.
Low coverage usually means the source text doesn’t contain the specs. Run Web Research to fetch specs from external sources first, then re-extract. Outcome: attributes populated even where your own data was thin.
Add normalization rules to the prompt (“Capitalize color names; express all dimensions in cm”). Re-run. Outcome: clean, uniform values that group correctly in filters and feeds.
Confirm the attribute exists on the product family, then enrich. If the value genuinely isn’t determinable from content, Web Research or a supplier import may be needed. Outcome: feed-ready products that pass marketplace validation.
Turn on review on Settings > AI review, or give one attribute its own rule in its AI review for this attribute section. AI values then wait in the Review inbox until you approve them. Track coverage in Data Quality Analytics. Outcome: confidence that extracted specs are accurate before they go live.

Web Research

Fetch specs from external sources when your own data is thin.

Data Quality Analytics

Track attribute coverage and catch suspicious values.

Managing Product Attributes

Set up options and switch on Let AI add new options.

Prompt Library

Build and organize your attribute prompts.