Auto-Drafts in Spark

The feature described in this article is currently in closed beta and available to a limited group of users. Some details may change before it becomes available to everyone.

Auto-Drafts is Spark's proactive reply helper. When an email lands in your inbox and looks like it needs an answer, Spark automatically prepares a draft response. All you need to do is review the draft, edit if needed, and send.

What is Auto-Drafts?

Auto-Drafts uses Spark +AI to detect incoming emails that require a reply and generate a response based on the conversation context, your writing style (if enabled), and your Personalization context (if enabled). The draft is saved to your Drafts folder and marked with a dedicated Auto-Draft indicator inside the opened email thread, so you can spot it instantly.

A few things to know:

• Auto-Drafts only appears under emails from real people — not newsletters, marketing, notifications, or receipts
• Like Auto-Labels, you can select which email accounts you enable Auto-Drafts for
• One draft is generated per eligible email
• Spark never sends Auto-Drafts on your behalf. You always review and hit Send yourself
• The feature works across all email account types Spark supports: Gmail, Outlook, iCloud, Yahoo, and any IMAP account

Prerequisites

• Auto-Drafts are currently available on desktop; once enabled on Desktop, they will be synced to mobile. For now, they will be displayed as regular drafts on mobile
• The Auto-Drafts feature is part of the Spark Pro plan
• Auto-Drafts work in individual inboxes only (Shared Inbox support is on the roadmap)
• Just like other Spark +AI features, it’s not available in China

How does it work? 

When Spark detects that an incoming email needs a reply, you have options that can be described as states:

• Base state — email looks like it needs a reply, Spark can automatically suggest a response based on the context of the thread. 
• Better state — if you're already using My Writing Style, Auto-Drafts can follow your unique writing style and tone. 
• Best state — we're also introducing personalization, which provides more context about your email habits, including your role and workflows.

How to spot Auto-Drafts in Spark

Auto-Drafts appear alongside your own drafts, with a few visual cues to help you tell them apart:

In your Inbox list, Auto-Drafts show up with the same "Draft" indicator as drafts you've written yourself — so your workflow stays familiar. 

In the Drafts folder, Auto-Drafts have a dedicated icon, making them easy to distinguish from your regular drafts at a glance. 

When you open an email with a pre-written draft, you'll see an "Auto-Draft" label at the top of the reply. 

What happens when you edit or share an Auto-Draft:

• If you edit it — it becomes a regular Draft, and is no longer marked as Auto-Draft in icons or labels.
• If you share it with your team — it becomes a Shared Draft, and is no longer marked as Auto-Draft in icons or labels.

How to enable and use Auto-Drafts

Auto-Drafts feature is available and can be configured in Spark Desktop, for each account separately. Once turned on, drafts created in Spark Desktop also sync to mobile, where they currently appear as regular drafts.

Follow these steps to turn Auto-Drafts on:

1. Open Spark Settings.
2. Go to Spark +AI > Inbox Automation.
3. Toggle Auto-Drafts on.
4. If Spark +AI isn't enabled yet, you'll be asked to do it.

Spark will start preparing drafts for emails that need a reply. You'll find them in two places: under the original message in your Inbox and in your Drafts folder.

Here’s what you can do with an Auto-Draft:

• Send: open the draft and press Send if it already says what you want
• Edit: start typing to refine the response. The Auto-Draft indicator disappears and it becomes a regular draft
• Discard: delete the draft 
• Share: invite a teammate the same way you would with any draft. The indicator changes to Shared Draft
• Report a bad draft: 

1. Open the Command Center (⌘K on your Mac/Ctrl+K on your Windows). 
2. Choose Give feedback on Auto-Draft.
3. Describe what was off. You can also share the thread with our team to help us troubleshoot.

You can also pause the feature temporarily or disable it: 

1. Go to Settings > Spark +AI.
2. Click Inbox Automation and toggle Auto-Drafts off. 
3. Alternatively, turn Spark +AI off entirely. Drafts already in your Drafts folder stay there until you send or delete them.

Privacy 

Auto-Drafts is part of Spark +AI and follows the same privacy rules as all Spark +AI features.

• Your data is not used to train AI models. When Spark generates an Auto-Draft, it sends the email thread, your writing style (if enabled; extracted from a sample of your sent emails), and your Personalization context (if enabled) to our AI provider, which returns the suggested response. None of this is used for training. For full details, see Spark's Privacy Policy.
• Spark doesn't learn from your edits or deletions automatically. If you rewrite a draft, delete it, or never send it, that behavior is not silently fed back into a model. The only time your drafts and edits are reviewed by our AI team is when you explicitly submit feedback through the Command Center.

Use case ideas

Auto-Drafts is built for the routine replies that take a few minutes here and a few minutes there but add up over a day. A few examples of where it helps:

• Quick acknowledgments: "Got it, thanks," "Will do," "Received, I'll review and get back to you", phrased the way you'd actually phrase them.
• Answering a question you already answered earlier in the thread: Spark uses the conversation context, so repeated questions get drafted responses pulled from what was said before.
• Follow-ups and confirmations: Short replies confirming next steps, approving a proposal, or saying yes to a request.
• Managing multiple accounts: If you're juggling a personal Gmail, a work Outlook, and an iCloud address, Auto-Drafts works across all of them so you don't have to switch contexts to keep up.
• Catching up after time away: Open your Drafts folder after a busy day or a vacation, and you'll find responses already prepared for the conversations that need one.

If you want to create more diverse drafts with a personal touch, we suggest generating a Personalization context — learn more about it here.