We present a study of the M-dwarf exoplanetary systems forthcoming from NASA's TESS mission. While the mission's footprint is too complex to be characterized by a single detection completeness, we extract ensemble completeness functions that recover the planet detections from previous work for stars between 3200 and 4000 K. We employ these completeness functions, together with a dual- population planet occurrence model that includes compact multiple planetary systems, to infer anew the planet yield. We predict both the number of M-dwarf planets likely from TESS and their system architectures. We report four main findings. First, TESS will likely detect more planets orbiting M dwarfs that previously predicted. Around stars with effective temperatures between 3200 and 4000 K, we predict that TESS will find 1274+/-241 planets orbiting 1026+/-182 stars, a 1.2-fold increase over previous predictions. Second, TESS will find two or more transiting planets around 20% of these host stars, a number similar to the multiplicity yield of NASA's Kepler mission. Third, TESS light curves in which one or more planets are detected will often contain transits of additional planets below the detection threshold of TESS. Among a typical set of 200 TESS hosts to one or more detected planets, 93+/-17 transiting planets will be missed. Transit follow-up efforts with the photometric sensitivity to detect an Earth or larger around a mid-M dwarf, even with very modest period completeness, will readily result in additional planet discoveries. Fourth, the strong preference of TESS for systems of compact multiples indicates that TESS planets will be dynamically cooler on average than Kepler planets, with 90% of TESS planets residing in orbits with e<0.15. We include both (1) a predicted sample of planets detected by TESS orbiting stars between 3200 and 4000 K, including additional nontransiting planets, or transiting and undetected planets orbiting the same star and (2) sample completeness functions for use by the community.