The TESS Data for Asteroseismology (T'DA) working group within TESS Asteroseismic Science Consortium (TASC) is responsible for processing the tens of millions of stars observed by NASA's Transiting Exoplanet Survey Satellite (TESS). In order to process this vast amount of data, we developed a machine learning framework to automatically classify the observed stars according to their stellar variability type. The framework aggregates the predictions from multiple distinct machine learning classifiers and combines their individual predictions into a global optimal classification by means of a metaclassifier. While this machine learning methodology and its resulting classifications are already of prime importance on their own for TESS science and future space missions, in this contribution we will specifically explore how the classifications can be coupled with spectroscopic information obtained from ground-based telescopes. This way, we will create a tool that is ideal for selecting the optimal targets for space missions such as ESA's upcoming PLATO (PLAnetary Transits and Oscillations of stars) mission, irrespective of whether it is a core or complementary science programme.