Safety & good practices

The AI Email Editor is designed to be a safe, focused assistant for building email templates. This page explains the guardrails that protect you, and how to get the best results.

What the assistant will and won't do

• Stays on topic. It helps with the email template you have open — explaining and editing it. It declines requests that are unrelated to building emails.
• Only touches the email you're editing. It works on the current template; another template you @mention is used as reference only and is never changed. It can't reach other clients' templates.
• Keeps content professional and safe. It writes in a professional collections tone, strips scripts out of HTML, requires secure (https://) image links, and won't add tracking pixels or unsafe code.
• Protects your data. It won't reveal internal details or invent customer data — placeholders are resolved per recipient at send time, not filled with real people's information in the editor.
• Never sends or saves on its own. Every change is a preview in the editor until you click Save (see Saving your changes).

If a request would break one of these rules, the assistant declines and, where it can, suggests a safe alternative.

Getting the best results

• Be specific about location. Say "the footer divider" or "the call-to-action button" so the assistant changes the block you mean.
• Name the exact content. Use an @mention for a template you want to compare against or borrow from, and name the placeholder you want inserted.
• Work in steps. Explain the template first, make a focused set of edits, then review.
• Review before saving. Always look over the changes in the editor — the assistant is a powerful helper, but you stay in control of what gets saved.

tip

Think of the assistant as a fast, knowledgeable pair of hands. You describe the intent and review the result; it does the clicking.