AI for Gmail: triage your inbox and draft replies, no busywork

You open Slack to forty unread emails you have not touched. You type "triage my inbox and tell me what actually needs me." Back comes a short list: three that need a reply today and why, two waiting on other people, a contract to review, and a note that the other thirty-five are newsletters and receipts already sorted. Two of the three replies are already drafted in your voice, waiting for you to read and send. You spent thirty seconds, not the usual hour of scrolling, and the inbox is now a to-do list instead of a wall.

Triage: sort the noise, surface what needs you

The problem with an inbox is not volume, it is that the four messages that matter are buried under the ninety that do not. AI triage flips that. Connected to Gmail, it reads the inbox and sorts it by what actually needs your attention, not by what arrived most recently.

• Needs you: a reply, a decision, or a deadline, surfaced with a one-line reason.
• Waiting: threads where the ball is in someone else’s court, so you stop re-reading them.
• Handled: receipts, newsletters, notifications, sorted and out of your way.

You get the short list of what matters and the confidence that nothing important is hiding underneath. Reading and sorting cannot break anything, so the AI does it freely and instantly.

Summarise the threads you keep putting off

The eighty-message thread you have been avoiding takes the AI seconds to flatten into a recap: what was decided, what is still open, who owes what, and the one question actually waiting on you. Instead of scrolling to the top and reconstructing the argument, you read four lines and know exactly where to step back in. The long threads stop being a tax on your morning.

Drafting in your voice, not a template

A generic AI reply is worse than no reply, because the recipient can tell. The fix is grounding the AI in your real material: how you have answered similar emails before, your tone, the actual thread it is replying to. An AI that reads how you write produces a draft that sounds like you, references the specific point, and lands the right next step, rather than "thank you for reaching out." You review it, tweak a line, and it is ready. Drafting is instant; you never start from a blank box.

Sending always waits for your approval

This is the line that matters. Drafting a reply is instant and reversible, you can always edit or bin a draft. Sending is neither: an email out is an email out, and the wrong one to the wrong person is a cost you cannot undo. So a trustworthy AI drafts freely and stops before the send, every time. It also extracts the tasks buried in your inbox, "you agreed to send the deck by Friday", into a list, but it does not act on them unattended. The pattern is consistent: the AI does all the work up to the irreversible step, then waits for your yes.

You stop living in the inbox

The quiet win is where this happens. You run all of it from Slack, not from Gmail. You ask for the triage, read the summaries, and approve the drafts in the channel you are already in, which means you are not pulled back into the inbox forty times a day to check it. The email still gets handled; you just stop living inside it. For most people that change, dipping in deliberately instead of refreshing compulsively, is worth more than the minutes saved on any single reply.

Running it through Echo

Echo connects to Gmail with one-click OAuth and runs from Slack. Ask it to triage and it returns the short list of what needs you, with the routine replies already drafted in your voice. Ask it to summarise a thread and it hands back the recap. Drafting and reading are instant; sending always waits for your approval, so nothing leaves your inbox unreviewed. Echo does not train on your email, encrypts your credentials, and does not charge per seat, so the whole team can use it without the bill climbing. The first $50 of work is free, which is enough to triage a real morning of mail. Start at /signup.