We cross-match 3XMM, WISE and FIRST/NVSS to create the largest-to-date mid-IR, X-ray, and radio (MIXR) sample of galaxies and AGN. We use MIXR to triage sources and efficiently and accurately pre-classify them as star-forming galaxies or AGN, and to highlight bias and shortcomings in current AGN sample selection methods, paving the way for the next generation of instruments. Our results highlight key questions in AGN science, such as the need for a re-definition of the radio-loud/radio-quiet classification, and our observed lack of correlation between the kinetic (jet) and radiative (luminosity) output in AGN, which has dramatic potential consequences on our current understanding of AGN accretion, variability and feedback.