Purpose. Within the SPACE project, a fully automated Dutch reading tutor, able to (1) track the child's reading progress and detect reading errors precisely by means of a speech recognizer; and (2) act as a fluent reading model and give adequate feedback by means of a speech synthesizer, is being developed. In order to explore the effectiveness of this tutor to improve a child's reading accuracy and fluency, an intervention study was carried out. Method. A pretest/posttest 4-week treatment with control group (no treatment) design was used. 10 pairs of reading disabled Flemish regular school children (grade 2-6) participated. Treatment consisted of non-repetitive reading sessions where the child could ask for phonic analysis help. When a child made an error but proceeded without asking for help, the speech synthesizer prompted phonic analysis feedback to the child in order to prevent the child from moving on before sounding out the word correctly. Results and Conclusions. In this paper we will discuss the effectiveness of the reading tutor mediated intervention sessions by comparing pre- and posttest reading skills for the experimental versus the control group. We will also have a closer look at how reading accuracy and fluency gradually improves with each reading session.