E-commerce: 40% less time to prepare campaigns

The problem

Every campaign meant assembling product cards and emails by hand. The bottleneck was not creative work — it was the repetitive production around it, and it set the pace of the whole calendar.

What we did

A content pipeline with templates and a QA gate, wired into the CMS: product descriptions built from the product's own attributes, email scenarios generated from the same source, checks running before publication.

Result

Time to prepare a campaign fell 40% over three months. Under NDA: the category and the trend, not the retailer.

We count the result as pages that entered the index and earn impressions on target queries — not as the number of cards published.

Method: where catalogue AI helps and where it hurts

Generation wins on volume of same-shaped text with checkable inputs. It loses wherever a fact is needed that is not in the data — and that is exactly where invented specifications appear.

From your own data only. A description is assembled from product attributes, not from the model's general knowledge. An invented compatibility claim costs more than the time it saved.

Uniqueness where the supplier text is identical. “How to choose”, “care”, “compatibility” sections are where you can differ, because everyone else ships the same vendor copy.

Checks before the index, not after. Duplicates, empty values and contradictions against the attributes get caught before publication, not after rankings drop.

A person on the sample. Proofreading thousands of cards is impossible; you check a sample plus every template the rest is generated from.

Services used on this project: AI automation, E-commerce growth.

Relative metrics only: these projects are under NDA, so the industry, the timeframe and the direction of change are published and the brand and absolute volumes are not. On a call we walk through the same structure applied to your market: baseline, work, expected metric.