Product Update: Better Drafts and Labels

At Lynda AI, we don't ship in a vacuum. Every feature is shaped by what our users experience day to day.

In this latest update, we focused on one thing: reducing inbox clutter and making replies feel more human and helpful. All of it came from repeated user feedback.

Here’s what’s new and why it matters.

No More Draft Overload

Until now, Lynda created a new draft for each message in a thread. That meant some users were seeing five or more drafts per conversation. It created clutter and made review more confusing than helpful.

(Multiple Drafts inside of a thread)

Now: Lynda updates her previous draft instead of creating a new one. She only replaces drafts she created, and she never touches drafts written by you.

This update-first approach keeps your inbox tidy and focused.

Replies Now Include More Quoted Context

When you hit "Reply all" in Gmail, it shows the conversation history, but the Gmail API doesn’t include that quoted content by default when creating a draft.

So we fixed that.

Now: Lynda includes the last 1 to 3 messages directly in the draft. It mirrors how a human would respond: your reply on top, and a clean quoted trail underneath it.

This makes reviews easier and the reply feel much more natural to recipients.

Better Labeling with Smarter Urgency Detection

Labeling (what some call tagging) is key to making sure the right messages rise to the top.

In this release, we updated how the classifier thinks about urgency:

• It now applies your rules first (not Gmail's labels).

• It uses tone and message meaning to set urgency.

• Urgency is only marked "urgent" when truly necessary, not just because it has a red label or looks important.

  
  

Messages are now more consistently sorted into the right categories, with fewer false alarms.

Start Your Free Trial

✏️ Better Drafts Using the "Conversation So Far"

For users on compact models like GPT-4.1-mini, we added an optional new feature behind the scenes.

Lynda can now see the last full message and up to three prior message summaries as part of her drafting context. This gives her a lightweight but highly useful way to understand where the conversation is heading and generate a reply that fits the flow of the thread.

What It Means for You

• Drafts now look more like something you would write

• Your threads stay clean and easy to review

• Fewer labels misfire

• And your inbox stays under control, quietly, without needing to configure a thing

Got feedback?

Keep it coming. If we hear the same thing a few times, it usually ends up shipped.