In recent years, cellular immunotherapy in B-cell malignancies has been driven by adoptive transfer of genetically engineered T cells expressing chimeric antigen receptors (CARs). CARs consist of a single chain variable fragment (scFv) of a monoclonal antibody, a spacer domain, a transmembrane domain, an intracellular signaling domain, and additional costimulatory domains. The bulk of clinical data available is on CD19-targeting CAR T cells for the treatment of B-cell acute lymphocytic leukemia (B-ALL), chronic lymphocytic leukemia, and B-cell non-Hodgkin lymphoma. Results so far have been promising with impressive rates and depth of remission especially among B-ALL patients. However, CAR T-cell therapy is a complex multi-step process, and clinical trials so far differ profoundly in CAR construct used, gene transfer method, composition of the cellular product, lymphodepletion, and CAR T-cell dose used. Randomized trials will be needed to conclusively evaluate the implications of these differences. The treatment concept is associated with significant neurotoxicity and potentially lethal cytokine release syndrome, both of which require specific management. Improvements in CAR design may help to overcome toxicity, the effects of an immunosuppressive microenvironment, and tumor escape by development of antigen-negative clones. This review will explain the mechanism of action, summarize the clinical experience with this treatment modality so far, and explore future developments in the field.