Generative AI has erased the tells defenders once relied on to spot phishing. The broken grammar, awkward phrasing, and obvious template errors are gone, replaced by fluent, personalized lures produced at massive scale. Worse, attackers can now clone a voice from seconds of audio and generate convincing video, enabling business email compromise and vishing that impersonate trusted executives in real time. This guide explains how to adapt defenses for that reality, shifting from spotting bad content to verifying context and behavior, hardening email and authentication, and building verification rituals that hold even when a request looks and sounds perfectly legitimate.