Multilingual product sheets for wine shops, importers and the online store: a spirits house selling abroad

Implementation example. This scenario describes a request we receive in this form. It is not a project delivered to a named client, and no results are claimed.

The context

A spirits house (cognac, armagnac, rum or whisky) sells in several countries. Its range includes permanent products, limited editions and seasonal gift sets. The marketing team is small: a marketing manager, sometimes an export manager, and agencies brought in as needed.

Each market has its own contacts:

importers and distributors, who ask for technical sheets and copy for their catalogs;
wine shops and online retailers, who want descriptions and visuals in their format;
the house's own online store, in two or three languages.

Product information lives in several places: a spreadsheet, PDF sheets, the ERP for logistics data, the inbox for the latest updates.

The day-to-day problem

Every new product or vintage sets off a chain of requests: rewrite the description, get it translated, format it for a given importer, update it on the website. Translations are ordered piecemeal, from different providers, and the vocabulary varies from one language to another. The tasting notes don't say the same thing from country to country.

Versions multiply. An importer publishes an old sheet with an alcohol content or bottle size that has since changed. The online store has an English version, but not a German one. The team spends its time looking for the right version instead of working on launches.

What the house wants

“I want a single source of truth for our products, and I want every market to have accurate sheets, in its language and its format, without chasing three providers every time we launch something new.”

The solution, step by step

1. Create a single source of product data. For each product, we gather the factual information: name, category, age or age statement, ABV, bottle size, aging, tasting notes approved by the cellar master, pairings, awards, logistics codes. Depending on the size of the range, this source is an existing PIM, a simple database or a structured spreadsheet. It is the only one that gets edited.

2. Build the glossary and the tone guide. Together with the team, we set the terms that must always be translated the same way (tasting vocabulary, appellations, cuvée names never to be translated) and the brand's tone. The glossary is approved once per language by someone from that market, ideally the importer.

3. Have AI write the sheets. From the source and the glossary, the AI produces for each product a short description, a long description and a sales pitch for trade buyers, in each language. It creates no information: everything it writes comes from the source sheet.

4. Adapt to the expected formats. The copy is generated in the requested formats: a PDF sheet for the importer, an import file for online retailers, content for the online store.

5. Have the market review. Each language is reviewed by someone in the country (importer, agent or trusted translator) before distribution. Their corrections are added to the glossary for the next sheets.

6. Follow each market's rules. Mandatory statements (in France, the Évin law and the health warning) and each country's own rules are checked with the importer before distribution. The AI does not validate compliance: it applies the rules it is given, and human review remains mandatory.

7. Update everything at once. When a piece of information changes in the source, the affected sheets are regenerated and the versions sent out are dated, so each partner knows which one to use.

What the AI does, what the house does

The AI writes and adapts the copy, applies the glossary and flags any information missing from the source. The house provides and approves the data, the cellar master approves the tasting notes, and the markets review their language.

Data and access required

the current technical sheets, even if incomplete;
the approved tasting notes;
existing translations considered good, which serve as the basis for the glossary;
the formats requested by the main importers and retailers;
one reviewer contact per language.

Project timeline

Weeks 1 and 2: diagnostic. Inventory of product sources, formats requested by each market and existing translations. This is the format of our audit (8 to 10 days).
Weeks 3 and 4: source and glossary. Structuring the product data, first glossary per language, tone guide.
Weeks 5 and 6: production. Generating sheets for a first batch of products, review by the markets, corrections added to the glossary.
Then: rollout. Extension to the full range and to new products as they launch.

Deliverables

a single, structured source of product data;
a glossary per language, approved by the markets;
a brand tone guide;
the range's product sheets in each language and each requested format;
a simple update procedure;
a short training session for the team.

Metrics to track

the time between the approval of a new product and the availability of its sheets in all languages;
the number of corrections requested by reviewers, which should drop as the glossary grows;
the number of products up to date on the online store and at the main retailers;
the time the team spends on product sheet requests.

These metrics are recorded at the start, then tracked at each launch. No results are promised in advance.

Limitations and prerequisites

Without a reviewer for a language, nothing is distributed: human review is part of the project.
The quality of the sheets depends on the quality of the source data. If tasting notes don't exist, they must first be written with the cellar master.
Each country's labeling and communication rules remain the responsibility of the house and its importers.

Indicative budget

The diagnostic corresponds to our audit, from €1,000 excl. VAT. The next phase is quoted on request, depending on the number of products, languages and formats.

Frequently asked questions

Does AI replace translators? No. It produces a first version consistent with the glossary. Someone from the market reviews each language before distribution.

Do we need a PIM? Not necessarily. For a short range, a well-structured database or spreadsheet is enough. We discuss it during the audit.

How do you stop the AI from inventing tasting notes? It only writes from the information approved in the source. Missing information is flagged, not invented.

Can we feed retailers' websites directly? When they accept an import file, yes. Otherwise, we provide them with sheets ready to integrate.

What next?

If every launch means back-and-forth with your importers and translators, the audit shows you in 8 to 10 days how to centralize your product data and sheets.

Client challenge