Treatment of high-grade serous ovarian cancer (HGSOC) remains challenging. Although HGSOC can potentially be responsive to immunotherapy owing to endogenous immunity at the molecular or T cell level, immunotherapy for this disease has fallen short of expectations to date. This Review proposes a working classification for HGSOC based on the presence or absence of intraepithelial T cells, and elaborates the putative mechanisms that give rise to such immunophenotypes. These differences might explain the failures of existing immunotherapies, and suggest that rational therapeutic approaches tailored to each immunophenotype might meet with improved success. In T cell-inflamed tumours, treatment could focus on mobilizing pre-existing immunity and strengthening the activation of T cells embedded in intraepithelial tumour myeloid niches. Conversely, in immune-excluded and immune-desert tumours, treatment could focus on restoring inflammation by reprogramming myeloid cells, stromal cells and vascular epithelial cells. Poly(ADP-ribose) polymerase (PARP) inhibitors, low-dose radiotherapy, epigenetic drugs and anti-angiogenesis therapy are among the tools available to restore T cell infiltration in HGSOC tumours and could be implemented in combination with vaccines and redirected T cells.