Automatic Speech Recognition systems use signal processing and machine learning in order to achieve speech transcriptions. Some analogies can be done with human speech recognition, but ASR use models that are much less complex than human brain. After a brief history of evolution of systems, the state of the art of ASR systems will be presented. The performance on various type of speech will be analyzed over various speech processing engines (from industrial and academic). One of the advantages of such system consists of the rapid production of transcripts that can raise the perspectives of analyses. Examples on automatization of speech task of verbal fluency of EVOLEX project will be given: logopedists and researchers benefit from advances with this kind of automatic treatments. Other advantage consist of the objectivity that automatic processing can give. For example, in C2SI project, in assessments for measuring the intelligibility of patients treated for ENT cancer, ASR can provide such advantage of a pool of speech therapist evaluations. Therapists can have subjective judgments of general speech intelligibility as they are used to ear the modifications of patient voice. Many aspects of voice can be analyzed with automatic processing tools: acoustics, prosody, comprehensibility. The main inconvenient of using ASR systems concerns the reliability and usage limits. Severe pathological voices infer very bad performance of automatic systems. The enhanced of recognition on such voice is not easy as state of the art systems necessitate thousand of hours of labeled speech in order to complete the learning process. We do not dispose of such amount of atypical voiced in order to improve the performances of ASR systems. Analyses must rely on extraction of cues in a more specific way. Speech Processing becomes to a certain degree of maturity. The use of such systems can transform some methodologies in voice treatment. The impact of these techniques do not have to be discarded and therapists can benefit to these evolutions.