Multilingual audit: 4 European markets in 1 audit

Prerequisites

• Active Agency+ plan (€599 excl. VAT/month). The multilingual feature is exclusive to this plan — see the pricing page.
• Client profile filled in with name, URL, sector. Languages are inferred from the markets but adjustable in the wizard.
• Sufficient credit budget. A multilingual audit consumes N times the credits of a single-language audit (where N = number of selected languages). The wizard shows an exact estimate before validation.

Step 1 : Open the launch wizard

From the client detail page (/client/{ID}), click the « Launch audit » button. The wizard opens in a 3-step modal: language selection, prompts review, models choice.

Step 2 : Select target languages

Tick FR, EN, ES, DE depending on the markets to cover. Up to 4 simultaneously by default. The platform validates that your plan allows the selected number of languages — Discovery is limited to 1, paid plans up to Agency+ allow 4.

Step 3 : Review localised prompts

The platform automatically rewrites prompts per market, not just translating them. A prompt asking « best French running shoes brand » becomes « best German running shoes brand » in German — the brand under audit stays the same, but the query matches what a local consumer would ask.

You can edit each localised prompt manually, language by language, before validation. The wizard shows a side-by-side preview.

Step 4 : Launch and monitor

Click « Launch ». Celery orchestrates the four executions in parallel. The real-time dashboard shows progress per market with a live counter (« 12/40 prompts done in French », « 8/40 in German », etc.). Average duration: 5–10 minutes for 4 languages × 30 prompts × 5 models.

Step 5 : Analyse the cross-market heatmap

Once finished, the dashboard displays a coloured matrix: one row per AI model, one column per language, green for strong presence, red for absence. You spot in two seconds the language × model pair needing attention. Switch to the « cards » view to filter by language, or to the « table » view to export a CSV.

For each underperforming language, a recommendations drawer opens with a contextualised action plan generated by Claude Haiku — specific to that market, not a generic recipe. See the dedicated article for examples.

Step 6 : Export the unified Premium PDF

From the dashboard, click « Generate PDF ». A single document is produced with: a common cover, a global summary on the first page, then four distinct chapters (one per market) with their scores, recommendations, detected competitors. Generation takes 5–10 minutes via Celery, you receive an email when ready. The PDF is fully white-label according to your agency settings.

Going further

• Article: 4 European markets in one audit (with concrete case study)
• AGS: the multi-judge anti-drift scoring engine
• Glossary: AI Share of Voice
• Pricing : Agency+ plan details