The National Tutoring Observatory (NTO) has introduced Sandpiper, an innovative open-source application designed to analyze large datasets of tutoring session transcripts. Co-developed with FreshCognate, Sandpiper employs agentic AI to annotate transcripts, identifying key tutoring strategies that can enhance student performance. Rene Kizilcec, associate professor at Cornell University, emphasized the need for efficient annotation methods, as traditional human coding is labor-intensive and costly. The NTO aims to create the largest open-access database of teaching data, collaborating with researchers from institutions like the Massachusetts Institute of Technology and Carnegie Mellon University. The app allows users to upload thousands of transcripts and apply AI models to analyze interactions at a granular level. Early feedback from pilot users indicates that Sandpiper significantly accelerates the research process, enabling higher precision studies. Future developments will enhance the app's capabilities, including data anonymization and analysis of non-verbal cues.