Descriptions copied from the manufacturer
A dozen stores carry the same text, so your product page loses to them in search.
Unique descriptions, meta titles and meta descriptions for hundreds of products, written from your store data. Specifications come only from your data, and you approve every description before it is saved.
A description sells the product and ranks the product page. We usually start when a store has one of these problems.
A dozen stores carry the same text, so your product page loses to them in search.
Customers do not find the answer they need and go back to Google or contact support.
Moving from another platform is the best moment to clean up descriptions and attributes.
A new delivery or supplier arrives and nobody has time to describe every item by hand.
The language model writes the text, but the facts, the structure and the decision to publish belong to the store. That is how we keep errors and duplicates down.
A description is more than the text under the photo. Search engines and customers read several elements at once.
Most problems with AI descriptions come from writing without data and without review. Each risk has a specific safeguard on our side.
| Risk | Where it comes from | How we prevent it |
|---|---|---|
| Invented specifications | The model fills missing data from general knowledge | Specifications from store data only; missing data puts the product on hold |
| Text like everyone else's | The manufacturer description rewritten in other words | A category template built around your customers, their questions and uses |
| Mass text written for robots | Keyword stuffing instead of answers for customers | Phrases used naturally, the description written for people first |
| Overwritten by an integration | BaseLinker or a supplier feed writes its own text on sync | Saved in the product's source of truth, with description sync switched off |
| Changes that cannot be undone | Thousands of descriptions saved without a backup | Saved in batches with a copy of the previous content |
Google's position on AI-assisted content: Google Search Central, February 2023. Text labelling: Regulation (EU) 2024/1689 (AI Act), Article 50, EUR-Lex.
The effect of new descriptions shows up in Google Search Console and GA4, usually a few weeks after they are saved.
How many product pages Google indexed before and after the descriptions changed.
How many different searches lead to product pages and how search traffic grows.
How the new meta titles and descriptions change the CTR of product pages.
Conversion of pages with new descriptions compared with categories not yet updated.
Net prices. Before you order, we prepare a free sample of a few descriptions from your store.
Main and short description, meta title and meta description, attributes from store data, approval and saving via API with a backup.
Smaller orders are billed at the minimum amount, which includes the category template and a sample for approval.
A draft Facebook and Instagram post for every new product, with a photo and the link in the first comment, for your approval.
Images for product pages are covered on the AI product images page. Wider business automation: automation and AI.
€1.55 net per product, with a minimum order of €260 net. The price covers the description, the meta title and meta description, attributes filled in from your store data, approval before saving and writing to the store via API. For catalogues of several thousand products we split the work into batches by category and start with the best sellers.
Not for using AI as such. Google states plainly that what counts is how helpful and original the content is, not how it was produced. What breaks its rules is mass-producing text mainly to manipulate search rankings. That is why every description is built on the data of a specific product and checked before publication.
Not with us. Specifications, dimensions, material, composition and compatibility come only from your store data, the manufacturer sheet or a spreadsheet you provide. The model writes text around that data and does not fill gaps from memory. If data is missing, the product goes on a list to complete instead of getting a guessed description.
A manufacturer description appears word for word in many stores, so Google usually shows one version, not necessarily yours. Our descriptions use the same facts but are written for your customers: the questions they ask, how the product is used, how variants compare and the phrases they search for. The facts stay the same, the text is yours.
The text obligation in Article 50 of the EU AI Act covers text published to inform the public on matters of public interest, and it does not apply to text reviewed by a person who takes editorial responsibility for it. A product description approved by the store usually falls outside it. Every description we write goes through approval, and any decision to label it is yours.
Via the PrestaShop or WooCommerce API, in batches, with a copy of the previous content, so every change can be rolled back. If you use BaseLinker or Apilo, we write descriptions to the system that is the source of truth for the product, so the integration does not overwrite them.
A new product in the store can get a description written to the same rules, plus a draft Facebook and Instagram post with a photo and the link in the first comment. The draft waits for your approval and nothing is published automatically. This service starts at €390 net to set up and €40 net per month.
See also: business process automation and AI, AI chatbot for websites, AI product images, online stores.
We will prepare a free sample of descriptions for a few products and price the whole catalogue. We reply within 24 hours.