Editing with chat
After the AI builds your email, you refine it the same way you created it — by talking to it. Ask for changes in plain English and watch the preview update. No menus, no markup.
What you can ask for
Conversational editing handles everything from tiny copy tweaks to structural changes. Common requests include:
- “Make the headline shorter and punchier.”
- “Change the button color to match my brand and label it Get started.”
- “Add a three-column feature section under the hero.”
- “Rewrite the body in a more casual tone.”
- “Swap the hero image for something brighter.”
Each instruction is applied to the current design, so you can iterate quickly without re-describing the whole email.
Checkpoints and revert
Every change you make is captured as a checkpoint. The Studio keeps a history of your versions, so if an edit goes the wrong way you can revert to any earlier point with a single click. This makes experimentation safe: try a bold idea, and if it doesn't work, roll it back instantly.
The Versions dropdown
The Versions button in the builder's top bar lists every checkpoint in your session, newest first, each labeled with the request that produced it and how long ago it was made. From the list you can:
- Preview any version in a pop-up before committing to it — the preview is read-only, so browsing history never changes your email.
- Rename a version you want to keep track of (for example, “Final hero layout”) using the pencil icon.
- Restore an earlier version. Your email returns to that state, and chat messages after that point are removed — the same behavior as reverting to a checkpoint from the chat.
The dropdown is available in both the AI Studio and the AI builder, and both surfaces share the same version history for a template.
Web research while you build
Sometimes the best email needs facts you don't have on hand — this season's trends, a competitor's launch, current statistics. With the Web research toggle in the chat composer switched on, the AI can search the public web and read pages mid-generation, then weave what it finds into your email.
- While it works you'll see live step chips such as “Searching the web: spring fashion trends” and “Reading vogue.com”, so research is never a black box.
- Each completed step stays in the conversation transcript, giving you a record of where the information came from.
- Research is capped per reply and per day, so a runaway request can't consume your plan.
Targeted vs. broad edits
You can be as specific or as open-ended as you like. Targeted edits (“change just the footer text”) keep the rest of the design untouched, while broad requests (“make the whole email feel more premium”) let the AI restyle multiple elements at once. If you need pixel-level control, switch to the code view and adjust the HTML directly.
Tips for great results
- Make one or two changes per message for predictable results.
- Reference elements by what they are (“the hero,” “the second button”).
- Run Remix for My Brand after the structure is settled to lock in your identity.
Conversational edits use AI tokens, so very long sessions consume more of your monthly allotment.
Frequently asked questions
Can I undo a change I don't like? Yes. Every edit creates a checkpoint, so you can revert to any earlier version with a single click — from the message's revert button or the Versions dropdown — experiment without worry.
Why did the AI change something I didn't ask about? Broad requests (“make it feel more premium”) intentionally touch multiple elements. If you want a surgical change, say exactly what to edit and what to leave alone.
Will editing break the responsive layout? No — the AI keeps the email responsive as it edits. Use the live preview to confirm how it looks on mobile and desktop, and ask for adjustments if anything needs tuning. When the design is right, lock in your identity with Remix for My Brand.