Escribo is a startup that originated at Johns Hopkins University and Porto Digital in Recife (PE, Brazil). All programs developed by the company are based on academic research and scientific evidence (meta-analytical and experimental studies). The instructional sequence, usage frequency, and basic program duration, consisting of 20 hours for each curriculum component (language and mathematics), were determined according to previous meta-analyses and validated through applied research in schools in Brazil, Panama, and Indonesia. One example is from our first randomized trial in schools serving students in poverty, where we identified during pretests that children lacked the prerequisite skills to benefit from the program. As a result, we expanded the program to include activities that foster those essential skills. Based on data collected by the games during experiments and correlating this data with standardized test results, we identified the best process metrics to use in learning reports. This evidence-based approach ensures that our product enhancements are informed by reliable data and contribute to improved learning outcomes. For more details, you can refer to our published research in Computers & Education Open: https://www.sciencedirect.com/science/article/pii/S2666557321000379 We also collect qualitative data from teachers' perceptions and school leaders' needs, which have been used to expand the program (e.g., to mathematics and reading fluency) and develop new features (e.g., adaptive tracks, mixed-methods reports, generating playful tips for families to use at home to foster the main skills needed by their children, etc.).