Voice Task Management: A Practical Guide

Voice task management means running your task list by speaking to it. You add tasks, change them, finish them and ask about them out loud, instead of typing and clicking. That sounds like one feature. In practice it covers two very different things, and most of the disappointment people have with "voice" tools comes from buying the first when they wanted the second.

This guide covers both, when speaking beats typing, how to set up a voice-first workflow, and what to check on privacy. We build Tarsen Voice, but most of this applies whatever tool you use.

What voice task management is

A task system does four jobs. It captures work so it stops living in your head. It sorts that work: priority, due date, which project. It helps you choose what to do now. And it records what happened: done, moved, handed to someone else.

Typing is the usual way to do all four. Voice task management swaps typing for speech in some or all of them. How many of the four it covers is the whole question.

Dictation vs talking to the board

Dictation turns speech into text. You press a button, speak, and the words appear where the cursor is, or as a new task with your sentence as the title. It covers capture and nothing else. After the words land, you're back to tapping: open the task, set the date, pick the project, drag it to the right place.

Talking to the board is a conversation with something that can read your tasks and change them. You say "move the invoice card to done" and the card moves. You ask "what's due this week?" and you get an answer.

The quickest test: can you ask it a question? If the microphone only ever produces text, it's dictation. If it can tell you what's overdue, it's reading your board, and it can probably change it too.

Both are useful. Dictation is fine if your only problem is getting thoughts down fast. But it leaves sorting, choosing and recording to your hands, and those three are where most of the time goes. We go further into the difference in what voice kanban actually means.

Why a Kanban board suits voice

A board is made of named things in named places, which is what speech handles well. "Move the billing migration to In Review" names the card, the action and the destination, with nothing left to guess. It works in reverse too: "three cards are due today: the invoice, the onboarding email and the pricing page bug" is easy to take in by ear.

When voice beats typing

When your hands or eyes are busy. Driving, cooking, walking. This is where tasks get lost, because the thought turns up when there's no keyboard.

When one sentence carries several fields. "Chase the invoice, due Friday, high priority" is three seconds of speech. Typed, it's a title field, a date picker and a dropdown. We cover this in more depth in add tasks by voice.

When you're updating, not creating. Finishing a task and then going to tell the tool you finished it is the job that makes boards go stale. "Done with the header fix, what's next?" takes two seconds and doesn't pull you out of what you're doing.

When you need to empty your head. Talking through everything on your mind is faster than typing it, and you don't edit yourself as much. A voice brain dump followed by sorting is a good way to start a week.

When you're planning the day. Talking through the board ("what's due? Push the report to Thursday, I'll take the two client cards") makes you decide, where dragging cards around lets you put decisions off. If you want a method for that conversation, see how to plan your day from a Kanban board.

When typing is the barrier. RSI, other accessibility needs, or just being someone who thinks better out loud.

When typing still wins

• Long, careful writing. A detailed card description or a spec reads better typed. Voice is for short instructions and quick capture.
• Exact text. Code, URLs, product names with odd spelling. Speech recognition gets these wrong more often than you'd like.
• Open-plan offices and shared spaces. Saying a one-line task quietly is no worse than a phone call, but nobody wants to narrate their whole backlog next to colleagues.
• Bulk rearranging. Dragging twenty cards into a new order is quicker by hand than describing it.

The realistic setup is mixed: voice for capture, updates and questions, the keyboard for writing and rearranging. What matters is that both land in the same place.

How to set up a voice-first workflow

"Voice-first" doesn't mean voice-only. It means speaking is the default way you put work in and keep the board up to date, and you reach for the keyboard when a task actually needs it. Here's a setup that holds up.

1. Pick one destination

Every spoken task should end up in the same system. If some go to a notes app, some to reminders and some to a text you send yourself, you've just built three inboxes to check. If your board is where work lives, voice should write to your board.

2. Name things so they're easy to say

You'll be saying your project and column names out loud, so make them short and different from each other. "Client work" and "Internal" are easy to say. "Q4 Client Deliverables (v2)" is not. The same goes for columns: "To Do", "In Progress", "In Review" and "Done" are easy to say and hard to mix up.

3. Say the details you already know

If you know the due date when you capture the task, say it then. Every field you leave for later is a small job you've given yourself. "New card in Client work: send the proposal, due Thursday, high priority" is one sentence and leaves nothing to tidy up.

4. Capture freely, filter later

Capturing a task isn't promising to do it. Get it onto the board and let your daily planning decide whether it matters. If you start judging every thought before you say it, you'll stop saying them.

5. Update the board as you finish

The biggest gain from voice isn't capture. It's keeping the board accurate. Make "done, what's next?" a reflex. A board that's always up to date is one you can trust, and one your team can read without asking you.

6. Ask before you scroll

When you sit down, ask the board what's due, what's blocked and what you're waiting on, instead of scrolling through every column. It's quicker, and it makes you look at the answer. Then keep the keyboard for long descriptions and fixing misheard names.

Privacy: what to check before you start talking

A voice tool hears more than a keyboard does, so ask a few direct questions before you use one for work.

• Is audio stored? Some tools keep recordings. Some keep nothing. Find out which.
• Where does the audio go? Most realtime voice features send audio to a speech or AI provider. Know which provider, and whether it goes straight from your device or through the vendor's servers.
• What happens to the transcript? How long is the text of a conversation kept, and who can read it? An admin? Your whole team? Only you?
• What can it change? Can voice delete things, change billing or remove people? Does it ask before doing anything that can't be undone?
• Whose permissions does it use? A voice assistant should see only what you can see. If you can't open a private project, it shouldn't be able to either.
• What's recorded on the board? When voice changes something, is that change attributed to you, and can a teammate see why it happened?

What to look for in a voice task manager

Beyond privacy, these questions separate dictation tools from ones you can run your work through:

• Can it act on the whole board, or only create tasks? Moving, reprioritising, assigning and commenting are where the time goes.
• Can it answer questions about your tasks?
• Does it know your projects, labels and teammates? "Put it in Client work for Sam" only works if it does.
• What does it do when you're vague? If there are two design cards, it should ask which one, not guess.
• What does it cost to talk? Realtime voice costs vendors money by the minute, so check the allowance, not just the plan price.

We compare the options in the best voice task manager apps. If you're on Trello and want to know how far voice gets you there, see Trello voice commands.

How Tarsen does it

Tarsen is a Kanban workspace, and Tarsen Voice is the talking-to-the-board kind, not dictation. Here's what it does and where it stops.

Starting it. Press ⌘J. Your browser asks for the microphone once. There's no app to install and no headset to buy. A workspace admin switches voice on for the workspace first, and AI has to be enabled there too. Setting up a new workspace offers it.

Talking. Speak normally. Tarsen works out that you've finished from what you said, not from how long you paused, so you can stop to think mid-sentence. You can also interrupt its reply.

What it can do. It can read the board, create and update cards, move them, assign them, label them, comment, work through checklists and lists, archive cards, and open any screen in the app. So "move the billing migration to In Review", "comment on TAR-142: the client approved the copy" and "what's blocked, and who am I waiting on?" all work. It can also type into a field you've selected when you ask it to.

What it asks first. Speech gets misheard, so two actions wait for a spoken yes: deleting a card and creating a project. Tarsen reads the name back and does nothing until you agree. Voice can open your billing page or member list, but it can't change either. There's no tool for that.

Whose permissions. Voice acts as you. It sees the projects you can see and nothing else, and the server runs the same permission checks the board uses. The AI model never touches the database directly. It asks for an action, and the server checks it.

Privacy. No audio is stored. Audio goes straight from your browser to OpenAI, because that's the only way to keep replies fast enough to interrupt. The text of a conversation is kept for 90 days, then deleted, and only the person who spoke can read it. Each change on the board carries the one sentence that caused it, so a teammate can see why a card moved without reading your whole conversation.

What it costs. Free includes 15 minutes for the whole workspace, once, so you can hear what it's like. Pro includes 30 minutes a month for each person, pooled across the workspace. Lifetime includes 15 minutes a month for the workspace. On Pro and Lifetime you can carry on with your own OpenAI key after that. Minutes reset on the 1st. Full details are in the voice docs.

Typing instead. Voice is built on Ask Tarsen, the assistant panel beside your board. If you'd rather type a question, the Ask Tarsen chat answers from your actual project and shows you a proposal before it creates cards. The keyboard can do everything voice can. For the wider picture of what AI does in Tarsen, see AI Kanban.

Where it stops. Voice needs a connection, because a session is live. It doesn't work offline.

Frequently Asked Questions

What is voice task management?

Running your task list by speaking: adding tasks, updating them, marking them done and asking about them, instead of typing and clicking. The useful versions go beyond dictation and let you talk to the board itself, so a spoken sentence can move a card or answer "what's due today?".

Is voice task management faster than typing?

For capture and short updates, usually. One sentence can set a title, due date and priority that would otherwise take several clicks. For long descriptions and exact text, typing is still better. The biggest gain is in moments where typing wasn't possible at all.

Is it safe to use voice for work tasks?

It depends on the tool. Check whether audio is stored, where it's sent, how long transcripts are kept and who can read them, and whether the assistant is limited to your own permissions. Tarsen stores no audio, keeps conversation text for 90 days, visible only to you, and acts with your permissions.

Try it on your own board

The quickest way to find out whether talking to your board suits you is to do it for a day. Try Tarsen Voice: the Free plan includes 15 minutes, enough to capture a few cards, move some others and ask the board what's due.