Modern factory standards, such as the Industrial Internet of Things (IIoT) and Industry 4.0 envision industrial processes that are evolving in time. An industrial process evolves following the effort to optimise its Key Performance Indicators (KPIs) as they are monitored by the underlying network. Optimisation is achieved with the addition, update or removal of assets within the industrial ecosystem, which, often, implies a reconfiguration of the underlying industrial control system. The reconfiguration of the control system of an industrial process, however, is a knowledge-intensive, time-consuming and error-prone task. This talk presents work in progress for a semantic framework based on session types for the automatic (re-)configuration of an industrial control system in the event of system evolution. Software components are described using session semantics that express the decentralised and distributed nature of a cyber-physical network. This implies a better utilisation of the resources with IIoT capabilities. Reasoning over session semantics identifies and synthesises the available network assets into a set of control loop session configurations. Given a of control loop session configuration, a control loop deployment module can automatically and seamlessly commit and deploy the necessary the control loop resources within the industrial network.