Overview of AI use in product texts

AI as a natural part of copywriting in 2026

How much of the product copy across 27 markets and 59 thousand products is covered by AI creation — split into copywriting, copyediting and translations.

Products total
59 219
across 27 markets
Unique texts (FC)
399 376
written by a human
AI texts total
66 899
AI generated + AI translations
Estimated savings
334 495 €
at €5 / unique text
Text types

Three ways texts are created today

Every product text carries one of three flags. They differ in the role of AI, the level of human oversight and the type of portfolio they suit.

Unique text

AI copywriting

The copywriter works with LLM models as a natural part of their job and is the guarantor of the final text's quality. An approach suited to specific, premium portfolios where standards can't be compromised.

  • Copywriter as the quality guarantor
  • LLM as an everyday working tool
  • Premium and specific portfolio segments
  • Flag: Unique text
AI text

AI copyediting

Product descriptions generated via LLM at larger scale, with copyeditor review according to the LQA standards of the given portfolio. It makes it possible to cover mass segments that previously stayed without any text.

  • Large-scale generation via LLM
  • Copyeditor review per portfolio LQA
  • Mass segments — previously without text coverage
  • Flag: AI text
Unique text · translation

AI translations

The translation is based exclusively on a verified Unique text — never on AI copyediting. The result is language versions of the CZ original that retain Unique-text status.

  • Source: always a verified Unique text (CZ original)
  • Never from AI copyediting
  • Language versions across markets
  • Flag: Unique text
Overall coverage

What the catalogue coverage is made of

Of 1 287 544 product impressions (product × country combinations where the product is actually shown), the text has the following origin:

Coverage across all markets

FC · AI · AI Tr are exclusive — each product has at most one text type
FC — human-written unique
AI — AI generated
AI Tr — AI translation
No text
Coverage by market type

What coverage looks like in individual segments

The market type indicates which price or distribution tier a product belongs to — luxury, mass, professional, etc. So far we deploy AI texts most prominently in the MASS segment.

Share of text types by market type

Stacked bar as a percentage of shown products
Coverage by category

Where AI texts help the most

Switch between the three catalogue levels. The lower the level, the more detailed the breakdown.

Share of text types by category

Stacked bar as a percentage of shown products
Coverage by country

Which markets have already gone AI

Poland, Czechia and Romania are the main pilots for AI generation. Most other countries have texts primarily from copywriters and so far only sporadic AI translations.

Share of text types by country

Sorted by absolute number of AI texts · RU excluded (data unavailable)
Savings calculation

How much we save versus unique texts

If we had to commission all 66 899 AI texts (AI + AI Tr) from a copywriter as unique copy, it would cost the following. Enter the exact price per text — the calculation updates live.

Default €5 / text — adjust to the real copywriter rate. Once you enter exact prices, the calculation captures the difference between what unique cost and what AI cost.

If you want net savings after subtracting AI costs, enter the price per AI text. Default 0.

Estimated total savings
334 495 €
334 495 €
Cost if it were unique
0 €
Actual cost of AI texts
54 266
AI texts
12 633
AI translations