Enriching Products with AI

Generate titles, descriptions, SEO metadata, attributes, alt text, and internal links for one product or your entire catalog while you keep working.

Select products, pick an enrichment type, configure it in a few clicks, and let AI do the rest. Every enrichment job runs in the background so you can keep working, and progress is tracked in the Process Tracker.

AI enrichment uses credits. Content types (titles, descriptions, SEO, attributes) each cost one credit per product. Translation, categorization, and image alt text have their own credit rates. Your credit balance is shown in the top navigation.

All 15 enrichment types

Type	What it does	Prompt needed
Auto-Fill	Detects every content gap across all 5 content types and fills them all in one run	Uses your project defaults
Bulk Enrich	Regenerates the fields you pick on every selected product, overwriting what is there	Uses your defaults, per field
Titles	Rewrites or generates product titles	Yes
Descriptions	Writes full product descriptions	Yes
Short Descriptions	Writes concise product summaries	Yes
SEO Title & Meta	Generates optimized meta title and meta description	Yes
Attributes	Extracts structured specs from existing content	Yes
AI Categorization	Assigns your existing categories based on product content	No
Internal Linking	Adds 3–5 contextual links to related pages inside each description	No
Search Intent	Rewrites descriptions to match real search queries from Search Console	No
Image Alt Text	Writes alt text for every product image that has none, using AI vision	No
Translation	Translates content into one or more target languages simultaneously	No
Linguistic Review	Scores translation quality and flags inconsistencies for one-click fixes	No
Quality Review	Scores each product across 12 data quality dimensions and surfaces fixes	No
Web Research	Searches the web for specs, pricing, and descriptions using your product identifiers	No

How to enrich products

1. Select products

   Check individual products, or use filters and select all matching results. You can also apply enrichment to your entire catalog.

2. Click Enrich with AI

   Click the Enrich with AI button in the toolbar. The enrichment modal opens.

3. Choose an enrichment type

   Pick what you want to generate. Content types require a prompt from your Prompt Library. The other types configure themselves.

4. Configure and start

   Set any options, prompt selection, language targets, confidence thresholds, and click Start Enrichment.

Finding the right action

The enrichment modal groups the actions so you do not have to scan a long list:

Group	What’s in it
Start here	One click, every field that needs work. The fastest way in if you are not sure what to pick.
Product content	Titles, descriptions, short descriptions, and attributes.
SEO & discovery	SEO title and meta, internal linking, search intent, image alt text.
Data & research	Web research and categorization: filling in facts rather than writing copy.
Languages & review	Translation, linguistic review, and quality review.

There is a search box at the top of the modal too. Type what you want to do and matching actions rise to the top. WISEPIM also preselects your default prompt for each field, so most runs are two clicks.

Content enrichment

The five content types each have a dedicated guide with prompt tips and examples: Titles, Descriptions, Short Descriptions, SEO Title & Meta, and Attributes. They all follow the same flow: select products, pick the type, choose a prompt, start. The AI uses the prompt combined with each product’s existing data and your Knowledge Library context.

Enriched content is stored in separate reviewed fields alongside your originals. Your source data is never overwritten.

Type	Output field
Titles	Reviewed Name
Descriptions	Enriched Description
Short Descriptions	Enriched Short Description
SEO Title & Meta	Meta Title, Meta Description
Attributes	Attribute values (extracted from existing content)

Set a default prompt for each content type in Settings → Enrich Prompts. It pre-selects the right prompt every time you open the modal, and is what Auto-Fill uses automatically.

Auto-Fill

Auto-Fill analyzes each product, finds every missing content field across all 5 content types, and fills them all in one run, using your project’s default prompts. No need to run five separate enrichments.

See Auto-Fill for the complete guide, including how to set the overwrite threshold to upgrade low-quality content alongside truly empty fields.

Bulk Enrich

Bulk Enrich regenerates the fields you pick across every selected product. This is the important difference: Auto-Fill fills the gaps, Bulk Enrich overwrites what is already there. Use it when the existing content is not good enough, not when it is missing.

Pick which fields to rewrite from descriptions, short descriptions, titles, SEO title and meta, and attributes. Each field can use your project default prompt, or you can override the prompt per field in the same screen, so a run can write descriptions with one prompt and titles with another.

Combine into fewer calls

Turn on Combine into fewer calls and titles, SEO title and meta, and short description are generated together in one call instead of three. Runs finish faster and cost fewer credits. Descriptions and attributes are big enough to keep their own call, so they are unaffected.

With Combine into fewer calls on, a product costs 1 credit if you picked one field and 2 credits if you picked several. The modal shows the estimated credit cost for the whole run before you start, so you always know what a run will cost.

Because Bulk Enrich overwrites, test it on 5 to 10 products first and read the output before you run it on the full catalog.

When credits run short

Credits are checked before a run starts and while it runs, so you never get a half-finished catalog with no explanation:

• If your credits only cover part of the selection, WISEPIM processes as many products as it can afford. The rest are skipped and marked in the task, so you can see exactly which products still need work.
• If you run out of credits mid-run, the task stops with a clear message telling you what happened, instead of silently failing.

Once you have credits again, open the task in the Process Tracker and choose Resume. Resume picks up only the products that were never processed, and it costs no extra credits because the original run already paid for them.

Resume finishes the remainder at no extra cost. Restart runs the whole selection again from scratch and is charged as a new run. When a partial run left products behind, Resume is almost always what you want.

AI Categorization

AI Categorization automatically assigns your existing categories to products based on their content.

Configure the confidence threshold (default 75%) and whether the AI determines the number of categories per product automatically or you set a maximum. Higher confidence means fewer but more accurate assignments. Advanced settings include the full confidence slider (0–100%) and a max categories control (1–10).

See AI Categorization for the full guide, including how to run a tiered pass that captures confident assignments first and only reviews the genuinely ambiguous products.

Internal Linking

Internal Linking reads each product’s description and adds 3–5 contextual links to related products and category pages. It prioritizes high-traffic pages using Search Console data.

See Internal Linking for the full guide, including how to pre-select target categories and when to expect the best results.

Search Intent

Search Intent uses your Google Search Console data to understand how real shoppers find each product, then rewrites the description to better match what they are actually looking for.

Requires an active Google Search Console integration. See Search Intent for the full guide, including how to identify the best candidate products.

Image Alt Text

Image Alt Text uses AI vision to analyze each product image and write descriptive alt text, up to 125 characters, for every image that has none. Existing alt text is never changed.

See Image Alt Text for the full guide, including accessibility and SEO benefits.

Translation

Translation converts product content into one or more target languages simultaneously. Configure which fields to translate, set a tone, add custom instructions, and save instruction templates for future runs.

For the full guide, including field selection, tone controls, strict vs. creative mode, and the terminology dictionary, see Translating Products.

Web Research

Web Research searches the web for product data using your product identifiers (EAN, UPC, model number, SKU, or brand plus name), cross-references multiple sources for accuracy, and puts the results in the Data Quality tab for you to review before applying.

For the full guide, including identifier selection and reviewing market price intelligence, see Web Research.

Linguistic Review and Quality Review

These two run from the same enrichment modal but open their own dedicated workspaces:

• Linguistic Review: scores your translated content across 6 criteria (completeness, terminology, locale quality, consistency, format, meta compliance) and offers one-click fixes for inconsistencies.
• AI Quality Review: reads each product like a demanding customer and scores it across 12 dimensions (completeness, content quality, image vs. data, pricing sanity, conversion readiness, and more), then hands you a ranked list of fixes.

Both require at least one target locale to be configured in Translation Settings before Linguistic Review appears in the modal.

Single product enrichment

Every enrichable field on the product detail page has an Enrich button. Click it to run enrichment on that field only, without selecting products in the main table. The same prompt selection and configuration flow applies, scoped to that one product and field.

Enrichable fields on the product page: name, description, short description, SEO title and meta description, attributes, and categories.

Using Content Logic with enrichment

Content Logic variables in your prompts give the AI dynamic, per-product context. A prompt like:

Write a description for {{product.name}} by {{product.brand}}.
Brand voice: {{brand.tagline}}. Target audience: {{category.target_audience}}.

resolves differently for each product, so the AI generates unique, contextual copy rather than variations on a single template.

See Content Logic for how to set up variables, and the individual scope pages for use-case examples.

Automating enrichment

All 15 enrichment types are available as Automation actions. Trigger them on a schedule, after imports, when quality scores drop below a threshold, or when products are updated.

Tips

• Start with 5–10 products. Test a prompt on a small batch and review the output before running it on your full catalog.
• Fill your Knowledge Library. The more context the AI has about your brand, audience, and products, the more consistent the output. See Knowledge Library.
• Use specific prompts. A prompt for “B2B industrial fastener descriptions” will outperform a generic “write a description” prompt every time. Create prompts per category or product type in the Prompt Library.
• Set defaults. Configure default prompts in Settings → Enrich Prompts so Auto-Fill works without any extra clicks and the right prompt is always preselected in the modal.

Related

Prompt Library

Build and organize the prompts that power content enrichment.

Prompt Lab

Test and refine prompts on real products before running them in bulk.

Knowledge Library

Give the AI your brand voice, audience, and product context.

Enrichment Impact

Measure the revenue lift AI enrichment generates across your catalog.

Teach WISEPIM your house style

Set up brand voice once so every AI-written description sounds like you.