Abstract—Modern technologies revolutionize human-machine interaction and task automation by successfully employing paradigms such as machine learning. In fact, deep learning, a machine learning technique based on artificial neural networks, has achieved commendable results, often exceeding human-level performance, in tasks that are impossible to solve by explicit programming, such as object detection and speech recognition. This project explores the application of such cutting-edge technologies in the context of electron microscopy in order to determine how they will affect the way people use microscopes in the future. More specifically, deep learning technologies were incorporated into an Intelligent Microscope (IM) prototype software system. The IM is a web-based system, powered by Artificial Intelligence (AI), that can be controlled through a voice interface, perform microscope operations, detect specific objects in microscope images, as well as interpret information about detected objects. This report presents the second development iteration of the IM, during which the system was extended to detect asbestos fibers, control multiple microscopes of different types using high-level voice commands, as well as perform image processing tasks such as smart denoising. The results of this work highlight the benefits of including AI technologies in microscopy systems. To be more specific, the achieved results include enhancing an existing asbestos detection workflow in terms of throughput, improving the usability of microscopes by means of a unified voice interface, and offering the ability to reconfigure a smart denoiser at will in order to optimize its performance according to the current operational settings of the microscope.