New 'Agent-Native' Analytics Platform Launches
A new analytics platform called Bonnard.dev has launched, billing itself as "agent-native." The platform is designed for rapid deployment of models and dashboards, claiming a 5-minute setup. It integrates directly with AI tools like Claude, Cursor, and Copilot, aiming to streamline marketing data pipelines.
Bonnard.dev's founder, Max Mealing, is a multi-time founder and self-described "expert generalist" focused on shipping products quickly and learning from failure. His stated philosophy is to "let your data be like water," emphasizing simplicity and rapid development cycles. The platform's core is a "semantic layer," which acts as a centralized translator for marketing data. It's designed to solve the common problem where different teams have conflicting reports because metrics like "conversion" or "customer acquisition cost" are defined differently across various tools. Being "agent-native" means it was built specifically for AI assistants to query directly, rather than for humans building dashboards. It uses a Model Context Protocol (MCP) server, which acts as a secure bridge between the data warehouse and AI models like Claude, allowing them to ask questions using governed, approved data. For a marketing analyst, this approach means an AI agent could be asked complex questions in natural language, like "Compare the ROI of our Q1 Google Ads campaigns to our organic social media efforts for the same period." The semantic layer ensures the agent uses the correct, pre-defined formulas for both ROI and acquisition cost. This technology aims to prevent situations where campaign performance data from a platform's dashboard conflicts with numbers in a company's internal database. By defining metrics once in a central place, all tools—from a Tableau dashboard to an AI agent—pull from the same source of truth. This shift impacts the skills entry-level analysts need, moving beyond just building reports in tools like Tableau. It places a greater emphasis on understanding how to model data and define metrics logically, often using SQL to query and validate the underlying data that these AI agents will ultimately use. The end goal is to enable analysts to move faster from a business question to a strategic insight without getting stuck on data cleanup and reconciliation. Instead of manually pulling and blending data from multiple ad platforms and a CRM to analyze campaign effectiveness, an agent can be tasked to do it, pulling from the pre-vetted semantic layer.
