Clearing an inbox by dictating the replies
· 6 min read
Most replies in a normal inbox are two to four sentences of ordinary prose, and that is the one thing dictation does better than fingers. The trick is not to dictate as you read. Triage the whole inbox first, decide what each message needs, then make one pass where you only speak. Done that way, a morning’s backlog takes about as long as reading it did.
Why email suits dictation better than documents do
Long-form writing by voice is a mixed experience, because you stop and restructure and change your mind, and restructuring by voice is slower than by keyboard.
Email does not work that way. By the time you have finished reading a message you usually know the answer, and the answer is short. There is no outline, no headings, no moving a paragraph three places up. You are converting a decision you have already made into three sentences. That is the narrow case where speaking beats typing by a wide margin, and it happens forty times a day.
The second reason is physical. An inbox is a long stretch at a keyboard producing very little text per keystroke, which is a bad ratio for your hands. Speaking most of the replies changes the shape of the morning.
Set it up once
macOS Dictation is in System Settings, Keyboard, Dictation.
- Turn it on and set the language.
- Set Microphone source explicitly to the microphone you actually use. Leaving it on Automatic means closing a laptop lid or connecting a display can quietly change which microphone is live.
- Choose the shortcut. Pressing Control twice is the default on most keyboards; keyboards with a microphone key can use that.
- Turn on Auto-punctuation. For email this matters more than it does elsewhere, because nobody wants to say “comma” forty times an hour.
- Click into any text field, press the shortcut, say a sentence, press it again to stop.
That is enough to start. The built-in shortcut is a toggle: dictation stays on until you stop it or it times out after silence. In an open office that means the microphone is sometimes still open while you talk to somebody at your desk, which is the main argument for a system where you hold a key while speaking and it stops the instant you let go.
The pass that actually clears an inbox
The mistake is answering messages in the order they appear, dictating each one as you meet it. You end up context-switching between reading, deciding and speaking, and you get the worst of all three.
Do it in two passes.
Pass one: triage, no replies. Go top to bottom with the keyboard only. Archive or delete anything that needs nothing. Flag the ones needing a real answer. Move the ones that are actually tasks into wherever tasks live. Do not compose anything. This pass is fast because every decision is one of four.
Pass two: dictate the flagged ones. Now you are doing one kind of work. Open the first flagged message, press Command-R to reply, dictate the answer, press Shift-Command-D to send. Next. The rhythm is what makes this quick: reply, speak, send, and no thinking about what to do with the message because you already decided.
Anything that turns out to need more than a few sentences goes back on the flag pile for later at the keyboard. Do not fight it in the middle of a pass.
Mail’s shortcuts are worth learning for this: Command-R replies, Shift-Command-R replies to all, Shift-Command-D sends, Shift-Command-U toggles read and unread. If you use Gmail in a browser, turn on its keyboard shortcuts in its own settings and learn the three you need; dictation works in a browser text field exactly as it does in a native app, because it goes to wherever the cursor is.
What to speak and what to type
Dictation is fast on prose and slow on everything else. Mixing them badly is where people conclude it does not work.
Speak: the body of the reply, the reasoning, the apology, the agreement, the question back.
Type: email addresses, URLs, prices, dates in numeric form, reference numbers, file names, and anybody’s name you are not certain the recognizer knows. Names are the single most common thing to go out wrong in a dictated email, and a wrong name in the first line undoes whatever time you saved. Dictating names, jargon and technical terms accurately covers building a vocabulary so more of these land.
Making dictated email not sound dictated
Spoken sentences run longer than written ones and they hedge more. Read three dictated replies in a row and you will hear it: everything starts with “so”, the sentences chain with “and”, and there are qualifiers nobody would type.
Four things fix most of it.
- Say the first sentence before you press the key. Composing while the microphone is open is what produces the “so, basically, I think what we should probably do” opening.
- Speak in whole sentences, then stop. Pauses become commas, and a stream of clauses becomes one enormous sentence.
- Cut the first three words of the reply after you finish. Nine times out of ten they are throat-clearing.
- Read it before you send it. Every time, for the whole message and not just the part you are worried about.
Some dictation software will do part of this for you. Voice Keyboard Pro has an optional Smart Rewrite that cleans filler words and shapes tone, which takes the “um” and “you know” out before the text lands. It runs on the text, not on your judgment: it will happily tidy a sentence that says the wrong thing, so step four stays.
Where dictating replies is the wrong call
Anything with legal or financial consequence. Not because dictation is inaccurate, but because the pace encourages sending. Type those, and leave them in Drafts for an hour.
Difficult messages. Declining, disagreeing, delivering bad news. These want slow sentences and the version you write at typing speed is usually better.
Rooms with other people in them. Dictating a reply about a colleague, out loud, in a shared office, is a category of accident worth avoiding entirely. If you work in an open space, keep dictation for the neutral messages and type the rest. Dictating in a noisy room: microphone and technique covers the acoustic side of the same problem.
What this actually saves
Be realistic about the gain. Dictation does not remove the reading, the deciding or the reviewing, and those are most of an inbox. What it removes is the typing, which on a forty-message morning is a genuine chunk of the time and most of the physical cost.
The other honest note: your first week will be slower, not faster. You are learning to compose out loud, and that is a skill with a learning curve like any other. Give it five days of real inbox before you judge it. How much faster dictation is than typing, and where it is not sets out how to measure the difference for yourself instead of guessing.
Questions
Does dictation work inside Mail’s reply window? Yes. Dictation types into whatever field has the cursor, so any app with a normal text area works, including Mail, browser-based mail and chat apps. The one thing to check is that you have clicked into the message body and not left the cursor in the subject line.
What happens if someone speaks to me mid-dictation? With the built-in toggle, their words get transcribed into your reply. This is the single most common embarrassment with dictated email. Stop dictation before you turn your head, or use software where the microphone is only open while you hold a key.
Is dictated email more error-prone than typed email? The errors are different rather than more numerous. Typing produces typos you can see; dictation produces real words in the wrong place, which spellcheck will not flag. That is why the read-before-send rule is stricter here than it is for typing.