Smart voice consists of analyzing the human voice in order to transform it into text. Everything goes through the voice, which is identified and then captured in sound frequencies. Then comes the analysis of these sound files using deep learning technologies linked to Artificial Intelligence. The data is translated into human language by speech recognition software. In order to solve the problems of syntax and comprehension the artificial intelligence (AI) technology is used, and in particular natural language understanding. The subsequent use of voice-to-text, deep learning, and language understanding makes it possible to obtain an optimal translation of the voice into data that can be interpreted by the machine, and from which the latter can then provide an adequate response.