Surface shape measurement is essential and ubiquitous in manufacturing. The dimensions of an object are often the most critical features when considering manufacturing tolerances, assembly with other parts and ultimately the functionality of the object. In addition, the range and complexity of shapes being manufactured has recently increased significantly as the adoption of additive manufacturing is embraced more widely. Shape measurement in manufacturing is often carried out using contact probing, which is slow, requires contact, and produces small data sets (or minimal surface coverage). To increase throughput and allow the adoption of digital approaches, industry needs to use faster optical methods that produce dense point clouds. In the work presented, we combine state-of-the art metrology and computer science to create robust autonomous 3D shape measurement systems, which enable measurement in a minimal amount of time and with efficient use of the available data. The novel and innovative parts of the research are related to the use of a priori data, using both material information and CAD models of an object to train artificial intelligence networks to respond to measurement challenges in real time. The research will enhance the measurement of shape in a manufacturing environment and reduce the need for highly skilled operators.