In practice, the concept of embedded AI as a very independent AI system with high real-time capability, efficiency and data protection is separated from that of edge AI and the development process is roughly outlined. The impacts of this technology on industrial applications (automation, white goods, automotive) are then shown and evaluated, starting with the use cases of a dish-smelling stove, predictive maintenance and stress detection in the face of a car driver. Finally, there is an outlook for the future.