Over the past two years, GCERF grantees in Bangladesh have tested various approaches to preventing violent extremism (PVE), and found that the following worked well: building cultural identity and promoting positive alternative narratives to strengthen social cohesion; introducing debate competitions in public and religious schools to enhance critical thinking; and organising frequent social mobilisation events to enhance community agency. However, activities such as large, non-interactive women’s gatherings, and using holy books and religious arguments as a counter-narrative to violent extremism, have proven to be ineffective in the Bangladesh context. This paper lays out some of these findings, which may help other P/CVE actors in Bangladesh for future programming.