Retool uses Loop to turn production data into AI roadmap decisions (www.braintrust.dev)

Retool, an enterprise app generation platform, has transformed its development process by integrating Braintrust's AI assistant, Loop, to make data-driven decisions for its AI roadmap. This innovative approach enables the team to leverage production logs to prioritize features based on actual user requests rather than assumptions. By semantically querying their production data, Retool gained insights into user needs, such as the high demand for multi-page support in their AI-powered assistant, Assist, which led them to prioritize the development of that feature. This shift toward an observability-first strategy underscores the importance of data in AI development. Retool's implementation of a classifier agent allows them to analyze user requests efficiently, calculating a "blast radius" metric to focus on issues that impact the most users significantly. Their systematic evaluation and feedback loop processes not only improve product quality but also build a shared understanding of priorities among team members. Retool's experience exemplifies how teams can better align engineering efforts with user needs through structured data analysis, ensuring that they respond to real demands while maintaining a high commitment to quality and continuous improvement.