The microstructure is seen as the primary information bearer that stores information about various process steps and also determines the mechanical and functional properties. Microstructure analysis is used daily: in quality assurance of existing, and research and development of new high-performance materials. Currently, microstructure analysis is still mostly done manually, or by simple computer-based approaches with very limited applicability. These approaches suffer from subjectivity and limited reproducibility, are time-consuming and cost-intense, do not access all relevant information from the microstructure, and therefore pose the bottleneck in microstructure-centered materials development and process optimization. Artificial Intelligence (AI) has the potential to revolutionize how microstructure analysis and materials-related research is accomplished. It offers tremendous benefits, like efficiency, savings, increased objectivity and reproducibility, and enabling new ways of microstructural analysis [1]. Ultimately, the automatic and efficient quantification of the microstructure also allows microstructural data to be collected in previously unattainable quantity and quality, which in turn forms the basis for establishing process-microstructure-property correlations and thus for microstructure-based and data-driven materials development and process control. In addition to some application examples for AI-based microstructure analysis and a guide on how to implement robust AI models for serial production, new opportunities in microstructure-based material design based on the results of AI-based microstructure analysis are outlined. The first results of a research project on the microstructure- and AI-based approximation of phase transformations will be presented. Long-term goals are a forward model to predict the achievable final microstructure depending on the initial microstructure, as well as an inverse model to determine the required initial state based on the desired final microstructure.