Datadog: SQL Is the New Bash for AI Agents

At MCP Night: The Holiday Special, Reilly Wood from Datadog demonstrated why SQL is a superior interface for AI agents handling large-scale data compared to traditional freeform commands or proprietary query languages. In a side-by-side demo, agents using SQL queries completed tasks more efficiently than those using Datadog's traditional tools, highlighting SQL's ability to manage context and perform efficient data aggregation. At Datadog's scale, SQL's structured nature not only enhances performance but also simplifies the AI agent's workflow by reducing the need for managing intermediate states or context windows. This approach aligns with broader industry trends, as both Datadog and Supabase recognize SQL's precision and reliability in modeling AI queries on internal systems, offering a structured, scalable solution for automated incident investigation and data analysis. The event underscored the idea that SQL, despite its constraints, provides the necessary precision for production systems, suggesting a clear preference for SQL over bash-like scripting in observability platforms.