Use case · WooCommerce product writer
Write 50 WooCommerce product descriptions overnight.
The agent reads your catalog, drafts descriptions for products that lack one, sets cross-sells and upsells based on real Woo data. Capability-scoped to <code>edit_products</code>.
The problem you actually have
Half your WooCommerce products have a 1-line description (“Blue cotton t-shirt, size M”) because you imported them from a CSV and never came back. Conversion suffers, AI search doesn’t understand the product, you know it’s costing you sales.
Generic AI writers churn out fluff that doesn’t match your brand voice. They don’t see your existing catalog, your style, or your real cross-sell logic.
You want: the agent reads existing well-written products to learn your style, drafts descriptions for the missing ones, also suggests cross-sells and upsells based on your actual category structure.
How the assistant solves it
You type the request in plain English inside wp-admin. The agent picks wp_recommend_products (plus wp_search_posts, wp_update_post as needed) and walks the steps below.
- You ask: “find my products with 1-line descriptions and write proper ones.”
- The assistant queries WooCommerce via the bridge to list products with short descriptions.
- It samples 5 products with good descriptions to learn your style.
- It drafts new descriptions for the flagged ones — same length, same tone, same level of detail.
- You review the markdown table, edit specific entries.
- The agent calls
wp_update_postfor each approved product (WooCommerce products arepost_type=product). - For cross-sells: “suggest cross-sells for these 12 products.” — agent uses
wp_recommend_productsto surface category-based recommendations.
Before vs after
| Without the assistant | With the assistant | |
|---|---|---|
| Time per product | 5-10 minutes | ~30 seconds per product including review |
| Style consistency | You drift across writing sessions | Agent learns from your existing well-written products |
| Cross-sells | You set them by hand, abandon halfway | wp_recommend_products uses category + tag adjacency |
| SEO | Stub descriptions hurt | Full descriptions with structured data ready |
Where it stops
- For products with very specific compliance requirements (food, cosmetics, regulated goods), the agent’s drafts can miss legal phrasing. Have a regulatory reviewer pass before publish.
- The bridge reads via WooCommerce REST namespace (
wc/v3); some custom WooCommerce extensions may not expose products consistently. Test with a small batch first. - Cross-sells are heuristic, not behavioral. For real behavioral cross-sells (customers who bought X also bought Y), you need WooCommerce Analytics + a recommendation plugin. The agent can configure those if installed.
FAQ
Does it work with WooCommerce extensions like Subscriptions or Bookings?
Yes for the description writing — those are still products at the post-type level. Subscription-specific metadata (billing intervals, etc.) requires the agent to use wp_set_post_meta with the right meta keys, which it can do but you have to guide.
Will it set prices?
It can but won’t by default — pricing is sensitive. You have to explicitly tell it in chat to set _regular_price meta on the product. We err on the side of safety.
Plan?
Admin or Bundle. Bundle is recommended for active WooCommerce stores because it includes the visitor chat too — your support traffic moves to the chatbot.
Stop shipping 1-line product pages.
Free plan, no card. Or read the full admin assistant page.