The automated analysis of microscopic images in materials science can be significantly improved through advanced deep learning segmentation models trained on ZEISS arivis Cloud. This platform enables subject matter experts, regardless of their AI expertise, to develop effective deep learning models, enhancing the analytical capabilities of material scientists. We present an easy-to-use AI-assisted annotation tool designed to streamline the training process for these models. Our solutions prioritize intuitive usability, facilitating the integration of sophisticated AI capabilities into existing workflows. The deep learning models trained on arivis Cloud can be integrated in our ZEN core image acquisition and analysis software, which provides a comprehensive solution that encompasses all stages of materials analysis—from microscope control and image acquisition to processing, analysis, and reporting. A primary focus of this presentation is the integration of deep learning instance segmentation models, which improve segmentation accuracy in complex scenarios involving overlapping or touching objects. This feature is particularly advantageous for grains analysis, where precise identification and classification of microstructural features is essential. We will present examples that demonstrate the effectiveness of instance segmentation models in real-world applications, highlighting their impact on the quality and efficiency of materials analysis.