Copper indium gallium selenide (CIGS) thin film photovoltaic devices are in the early stages of large-scale commercialization. Their high performance, uniformity, reliability, and a low carbon footprint make them an attractive alternative to standard silicon solar cells. Due to the complex processing required and the associated manufacturing costs, reliable in-line quality control technology is needed. By identifying defective cells early in production, faulty batches can be excluded from further processing, saving resources and costs. We show that micro-Raman spectroscopy (RS) and hyper-spectral imaging (HSI) are powerful tools for quality control and process improvement. Distinctive features in the Raman spectra allow the estimation of the copper to gallium plus indium (CGI) ratio, which is an important criterion for the cell’s efficiency. With HSI in the visible and near infrared range (VNIR) and the near-infrared spectral range (NIR) in combination with machine learning techniques, the layer thickness and CGI ratio are accurately predicted.