In the sense of the process-structure-property chain, to add value to the development of new advanced materials, materials design approaches must be tailored to support downstream optimal processing approaches. In this contribution, we combine two recently developed machine learning-based approaches, one identifies near-optimal material microstructures for desired properties and the other aims to guide manufacturing processes to produce identified microstructures. It is worth noting that, both the identification of near-optimal microstructures and the search for optimal processing routes, are ill-posed inverse problems and are therefore, typically non-unique (i.e., more than one solution exists). This is a challenging issue for solution approaches but also offers an advantage which is leveraged in the used machine learning-based approaches to guide the underlying production process efficiently to produce the best reachable microstructures. In this contribution, the approaches presented are validated at the example of a simulated metal forming process, aiming to optimize crystallographic texture.