In spite of its relevance, the Thermally Pulsing Asymptotic Giant Branch (TP-AGB) phase is one of the most uncertain phases of stellar evolution, and a major source of disagreement between the results of different population synthesis models of galaxies. I will briefly review the existing literature on the subject, and recall the basic prescriptions that have been used to fix the contribution of TP-AGB stars to the integrated light of stellar populations. The simplicity of these prescriptions greatly contrasts with the richness of details provided by present-day databases of AGB stars in the Magellanic Clouds, which are now being extended to other nearby galaxies. I will present the first results of an ongoing study aimed at simulating photometry, chemistry, pulsation, mass loss, dust properties of AGB star populations in resolved and un-resolved galaxies. We test our predictions against observations from various surveys of the Magellanic Clouds (DENIS, 2MASS, OGLE, MACHO, Spitzer, and AKARI). I will discuss the implications and outline the plan of future developments.