Learning to read is one of the fundamental skills children acquire at school, and developing fluent decoding skills is fundamental to become a good comprehender. Fluent reading includes accurate, rapid, and expressive reading and objectively gaining in-depth diagnostic insights into these skills in a school-setting remains difficult. Objective in-class diagnostic testing with individualised guidance through personalised learning-to-read trajectories would be possible if reading skills could be assessed through an automatised speech-enabled digital system that (a) analyses speech in reading aloud tasks and (b) models the gathered data and provides information on how to improve reading. This can be achieved by combining Automatic Speech Recognition (ASR), speech diagnostics and learning analytics into an innovative, integrated approach to reading diagnostics. The interdisciplinary research conducted in this project aims to investigate how speech diagnostic output can be obtained from ASR and how these diagnostics can be converted into recommendations on personalising learning-to-read trajectories. To showcase these lines of research, an educational app will be developed in which reading exercises for young primary school children are analysed through ASR technology, and learning analytics are used to model ASR data and provide recommendations on how to foster a child’s reading development. This novel interdisciplinary approach will not only provide new scientific insights into how accuracy, fluency and expressive reading develop over time, thus increasing our understanding of how children learn to read, but it will also put this knowledge into practice to evaluate its impact on the process of learning to read.