The chromatin-associated protein WDR5 (WD repeat domain 5) is an essential cofactor for MYC and a conserved regulator of ribosome protein gene transcription. It is also a high-profile target for anti-cancer drug discovery, with proposed utility against both solid and hematological malignancies. We have previously discovered potent dihydroisoquinolinone-based WDR5 WIN-site inhibitors with demonstrated efficacy and safety in animal models. In this study, we sought to optimize the bicyclic core to discover a novel series of WDR5 WIN-site inhibitors with improved potency and physicochemical properties. We identified the 3,4-dihydrobenzo[f][1,4]oxazepin-5(2H)-one core as an alternative scaffold for potent WDR5 inhibitors. Additionally, we used X-ray structural analysis to design partially saturated bicyclic P7 units. These benzoxazepinone-based inhibitors exhibited increased cellular potency and selectivity and favorable physicochemical properties compared to our best-in-class dihydroisoquinolinone-based counterparts. This study opens avenues to discover more advanced WDR5 WIN-site inhibitors and supports their development as novel anti-cancer therapeutics.