There are two ways to get from a blank canvas to a finished email, and they trade off along different axes.
The manual drag-and-drop approach is the one every email builder has trained you on: pick blocks from a palette, drop them on the canvas, click into each one, and configure it through a properties panel. Precise, predictable, and fully in your hands.
The AI-assisted approach flips the input. Instead of clicking, you describe the email in plain language to Claude, ChatGPT, or another assistant connected through the Model Context Protocol and the assistant builds it by calling the same operations the visual editor exposes. You describe, it assembles, you review.
Neither replaces the other. The useful question is which to reach for, and when and the answer is almost always both, in the same session.
