BlaBlaCar rebuilds its translation workflow with AI and reports around 27% lower translation costs

The problem

BlaBlaCar runs in more than 30 markets and languages. Every product change, campaign or feature update has to be translated and adapted quickly. The old process had many manual steps, coordination across tools and back-and-forth to keep quality consistent. The team wanted to scale without raising complexity, cost or turnaround time.

What they built

The localization team redesigned parts of its pipeline to use AI for translation, quality checks and automation. The content team says the new setup takes over a large part of the repetitive work, with people still deciding on quality and language.

The localization manager's own post describes the tool. It started as a Gemini call from inside a Google Sheet. It grew into a Sheet with a glossary, style guide, examples of approved translations and a second set of prompts that checked the output. In Q3 2025 it became a script connected to Phrase, the company's translation management system.

How they did it

A person uploads new content to Phrase and adds two tags: one to request AI translation and one to pick the languages. A script then collects the content and the company's glossary, style guides and examples, translates the text and runs a quality-check prompt. It writes the result back to Phrase. Linguists find the text already translated and checked, and do the final review.

The manager built the script in August 2025 with AI as a coding partner, which took dozens of prompts, refinements and tests.

Results

The content team's post says that across 2025, AI-assisted workflows brought translation costs down by around 27% and improved average turnaround times by 22%. The localization manager's November 2025 post reports 24% savings on translation costs for the automated system and over 6 hours of manual work saved each month. It reports earlier translation savings of 11% (the first spreadsheet formula) and 22% (the larger spreadsheet tool). That 22% is a cost saving at one stage of the tool, not the 22% turnaround figure.

Localization experts now spend more time orchestrating workflows, defining language standards and checking quality across markets, and less time executing translations.

What to copy

Begin with a small tool, put your glossary and style rules into the prompts, add an automated check, and keep linguists as reviewers. Report savings at each stage so the business case grows with the tool.