This study investigated the appropriate ways in offering immediate automated writing feedback within the framework of process-based writing pedagogy by comparing relative effects of two different automated writing evaluation (AWE) system application types on improving writing performance. The experiment took an initiative step in elucidating at what point in the process-based writing stages AWE feedback is best to be served. The research is conducted to confirm whether providing an instant language-related feedback whenever EFL students call for will either interfere the development of content indeed -- as it has been expected by process-based writing approach, or bring improvement in students' writing. Two application types, namely non-continuous feedback (NCF) and continuous feedback (CF) group -- are differentiated in terms of in which point of the writing stages students are enabled to get access to the AWE system. With the purpose of the study, a total of 20 students participated. The findings revealed that CF group did not receive language-related AWE feedback to the point of distracting the development of their content. Furthermore, CF group significantly outperformed NCF group in overall writing product, especially on the dimension of grammar and content. Students also expressed a positive attitude toward receiving instant language-related feedback via AWE system.