Triple-negative breast cancer (TNBC) is a distinct subtype characterized by the absence of estrogen, progesterone, and human epidermal growth factor receptor-2 expression. Representing 15–20% of newly diagnosed breast cancer cases, TNBC carries a 40% mortality rate within the first 5 years post diagnosis. This underscores the urgency in enhancing TNBC chemoresistance, one trustworthy option being the use of natural compounds as chemosensitizers. The main objective of this study was to improve the sensitivity of TNBC cells to docetaxel by combining it with cytotoxic natural compounds and to elucidate the precise molecular mechanisms of action using a cellular and metabolomic approach. In pursuit of this, we conducted cytotoxicity profiling for docetaxel, genistein, leontopodic acid, and kaempferol individually and in dual combinations on a panel of TNBC cell lines (HS578T, MDA-MB-231, MDA-MB-462), using the MTT assay. The results indicated that the genistein–docetaxel combination exhibited the most potent cytotoxic effects across all cell lines. For this combination, we investigated the biochemical mechanisms involved using a gas chromatography–mass spectrometry (GC-MS) metabolomics approach. The analysis identified 69 intracellular metabolites, with 17 quantified relative to internal standards. These metabolites belong to different chemical classes, including phosphorylated amino acids, carbohydrates, fatty acids, or nucleotides, and are key intermediates in different metabolic pathways. Quantitative enrichment analysis revealed distinct pathway alterations for each cell line, encompassing beta-alanine metabolism, nicotinamide and nicotinate metabolism, purine and pyrimidine base metabolism, as well as lysine degradation and biotin metabolism. In conclusion, the combination of genistein and docetaxel demonstrates a potent chemosensitizing effect in TNBC cells, inducing cytotoxicity through diverse metabolic pathways. This underscores the considerable potential of this combination in addressing the TNBC subtype. Acknowledgments: This work was supported by a grant of the Ministry of Research, Innovation and Digitization, CNCS— UEFISCDI, project number PN-III-P1-1.1-PD-2021-0093, within PNCDI III, awarded to Alina Uifălean.