Using Product Content

Last updated: August 21, 2026

Overview

The Product Content skill (/product-content) rewrites product page copy — titles, descriptions, bullets, specs, call-to-action labels, and imagery callouts — in your store’s own voice. It reads your live product pages, combines what it finds with your Noibu behavioural data, and recommends specific changes as a before-and-after view you can edit and approve field by field.

By default it works toward improving add-to-cart rate: it finds product pages that get real traffic but are not moving shoppers, narrows to the ones where the copy is a likely contributor, and recommends changes to those. You can also give it a direction instead — a tone shift, a campaign alignment, or a brand-voice clean-up — and it applies that on top of your existing voice.

This skill is not included as part of the Noibu plugin. It can be installed from the Skill Store.


When to use it

  • When product pages are getting traffic but not add-to-carts, and you suspect the copy is part of the reason

  • When aligning a collection with a campaign or a seasonal push

  • When older listings have drifted from how the brand sounds today

  • After running Product Analysis, when it pointed at a merchandising or content problem ( /product-analysis diagnoses and /product-content executes changes)

Requirements

  • A browser extension for your AI tool, such as the Claude extension for Chrome, recommended. Many storefronts render product content in the browser, and the extension is the most reliable way for your AI tool to read the live page.

  • The Shopify connector, to publish changes directly. Without a connector you can still review everything and export the changes to apply manually.

How to trigger it

There are three ways to run it, depending on how specific you want to be.

  • Broad sweep. Type /product-content or ask “improve the product content on my site.” The skill finds the product pages most worth fixing and shows you why before recommending changes.

  • Point it at products. Name the products, SKUs, or a collection — “audit the copy on these products”, or “why isn’t this product page converting?”

  • Give it a direction. Describe the change you already have in mind — “make the content for products in the Spring collection more uplifting”, or “rewrite these to sound more premium.” You can combine this with a named product set.

What you get

A Recommended changes card listing each product with a table of field changes: the current copy beside the proposed copy, with a short rationale per product. From there you can:

  • Edit any proposed copy inline, or ask your AI tool in chat to revise it

  • Keep or drop changes individually, or accept everything for a product at once

  • Export the changes in a range of document and spreadsheet formats, to hand to someone else

  • Publish the kept changes straight to your store, if you have a supported ecommerce connector

Publishing and measuring the result

Product Content publishes product data — titles, descriptions, bullets, images, and metafields — through your ecommerce platform’s own product API, not through GitHub. That data lives in your platform’s database rather than in your theme code, so publishing is immediate once you click and no pull request is involved.

When you publish, the skill also records the change as a release event in Noibu. You can see it in the Noibu console or from your AI tool.

After a successful publish, the skill offers to run on a recurring basis. Each scheduled run opens with an add-to-cart movement report — how your site has moved since the last run, and the before-and-after rate for each product it updated — before recommending a fresh set of changes.