This work presents the realization of a maintenance device assisted by artificial intelligence. This system makes it possible to improve the supervision of the network, the analysis of technical feedback and their correlation to reduce incident resolution times and better plan or initiate interventions. It is based on the ZABBIX package for monitoring, the Alexa package for conversational agent as well as on some Artificial Intelligence bricks such as AVS, IoT, AWS... It is added to the list of possible solutions and acceleration levers for the growth of operators by reducing the time to detect faults in the network.