BlaBlaCar identified a significant disconnect between their Software Engineers (SWE) and Data Analysts (DA). While both groups have valuable skills, engineers often feel intimidated by the complex tools needed for data analysis, leading to a reliance on the data team for basic insights. This creates a bottleneck where analysts juggle numerous quick questions while also trying to maintain data integrity. The solution they propose is the “BlaBlaCar Data Copilot,” which aims to empower engineers to conduct their own data analyses and reduce the dependency on analysts. The concept of "Shifting Left," borrowed from DevOps, is central to this initiative. Instead of analysts catching data quality issues post-production, engineers would validate data during feature releases. This transition not only speeds up the process but also transforms analysts from search engines into mentors. By operationalizing the Data Mesh, BlaBlaCar encourages domain teams to take ownership of their data products, prioritizing data quality as a preventative measure rather than a reactive one. The Data Copilot leverages Large Language Models (LLMs) to bridge the gap between software engineering and data analysis. It provides curated queries tailored to BlaBlaCar’s specific business needs, ensuring that engineers don’t just work with generic SQL. Each analysis includes a Data Health Card that checks for potential issues, allowing engineers to learn from mistakes without being hindered. Additionally, insights are treated as transparent artifacts, with generated code for analyses that can be easily modified. This cultural shift aims to make data analysis a collaborative and iterative process, rather than a one-off task.