Short answer: Xero has no built-in dictation, but because it runs in a browser you can dictate into any Xero text field with a system-wide voice tool. On Mac, click the field, hold your hotkey and speak. On iPhone, switch to a voice keyboard in the Xero app and tap the mic.
Xero is a writing application disguised as an accounting application. Between the invoice line descriptions, the transaction memos, the reconciliation notes, the contact records, the file annotations and the emails you send chasing payment, a bookkeeper who runs a handful of clients through Xero can easily spend more of the week typing sentences than entering numbers. Nobody talks about this part of the job in the software demos, but it is where the hours actually go.
The reason it goes unnoticed is that each individual piece of writing is small. A transaction memo is one line. A contact note is two sentences. A reconciliation comment is a fragment. But you write hundreds of them a month, each one requiring you to move your hands from the numeric keypad to the letter keys, compose a sentence, and move back. That context switch is the tax, and it compounds.
Voice removes the context switch. You keep your eyes on the bank feed and your hands wherever they are, hold a key, say the memo out loud, and it lands in the field. This guide covers how to set that up in Xero on both Mac and iPhone, which fields are genuinely worth dictating, and, just as importantly, which ones you should keep typing.
Why Xero has no dictation of its own
Xero is a browser-based application. It runs inside Chrome, Safari, Firefox or Edge, and its input fields are ordinary web form fields. Xero does not ship a microphone button, does not offer a voice entry mode, and there is no dictation plugin in the Xero App Store that types into arbitrary fields for you. That sounds like bad news, and it is exactly the opposite.
Because Xero uses standard browser text fields, anything that can type into your Mac at the cursor position can type into Xero. There is no integration to install, no OAuth connection to approve, no third-party app touching your financial data through the Xero API. The voice tool never talks to Xero at all. It simply produces keystrokes, and Xero receives them the same way it receives keystrokes from your keyboard.
This is a meaningfully better privacy and IT posture than an integration would be. Nothing gets read out of your organisation's ledger. Nothing gets write access to your clients' books. The tool converts your speech to text and inserts it wherever your cursor happens to be, which today is a Xero field and tomorrow might be an email.
Setting it up on Mac
Voice Keyboard Pro sits in the Mac menu bar and works system-wide. The setup for Xero is the same as the setup for anything else, because there is no Xero-specific step:
- Install the Mac app and grant it microphone and accessibility permission when prompted. Accessibility is what lets it insert text at your cursor rather than only into its own window.
- Pick a hold-to-talk hotkey you can reach without looking. Something you do not use in your browser or in Xero's own keyboard shortcuts.
- Open Xero, click into any text field, hold the hotkey, speak a sentence, and release. The text appears where your cursor is.
One setup detail worth attention: Xero has its own keyboard shortcuts, and browsers have many more. Before you commit to a hotkey, open Xero and try it in a few screens to make sure it does not trigger a search bar, a new tab, or a Xero navigation shortcut. A modifier-key combination you never otherwise press is the safest choice. If your hotkey seems to do nothing in the browser but works in Notes, a shortcut collision is almost always the reason.
On iPhone and iPad
The Xero mobile app is where a lot of the small, awkward writing happens. You photograph a receipt in a car park and need to describe it. You take a call from a client and want the note in the contact record before you forget it. Typing that on a phone keyboard is the slowest writing anyone does.
Install the Voice Keyboard Pro keyboard, add it in Settings under General, Keyboard, Keyboards, and allow Full Access so it can reach the microphone. Then in the Xero app, tap into any field, press the globe key to switch keyboards, tap the mic, and talk. It works in the Xero app the same way it works in Mail or Messages, because iOS treats the keyboard as a system component rather than an app-specific feature.
The Xero fields worth dictating
Invoice line descriptions
This is the single highest-value field in Xero for voice, and it is the one most people under-write. A line item that says "Consulting" tells the client nothing and generates a query email three weeks later. A line item that says "Strategy workshop facilitation, two half-day sessions with the leadership team, including preparation and the written summary" answers the question before it is asked and gets invoices paid faster.
The reason most invoices contain the first version and not the second is that the second version takes forty seconds to type and you have eleven of them to do. Spoken, it takes eight seconds. The economics of being descriptive change completely when the marginal cost of a good description drops that far.
Click the description field, hold the hotkey, and describe the work the way you would describe it to the client on the phone. That conversational register is exactly the right one for an invoice line.
Transaction memos and reference notes
When you are working through a bank feed and something does not fit a clean category, the memo field is where future-you gets rescued. "Reimbursement to the client for the courier they paid on our behalf in March, invoice to follow in April" is the kind of note that saves twenty minutes at year end and takes five seconds to say.
Dictating memos also changes what you write. Typed memos get truncated to a couple of words because typing is friction. Spoken memos come out as full sentences with the actual reasoning in them, which is what makes them useful when someone reads them nine months later.
Bank reconciliation comments and discussion
Reconciliation is a queue. You go item by item, and any item you cannot resolve immediately needs a note explaining why so you or the client can come back to it. Because you are in a rhythm, the temptation to leave a note that says "check" is enormous. Voice lets you leave "check whether this duplicate charge was refunded, the merchant's statement shows two attempts on the same day" without breaking the rhythm at all.
Contact notes
The notes on a Xero contact record are where client context lives: who signs off, which entity to bill, what their payment behaviour is like, what was agreed on the last call. Contacts records everywhere are chronically under-populated for the same reason invoice lines are terse, and the same fix applies. Speak the note right after the call, while you still remember the specifics.
Emails you send from and around Xero
Payment reminders, statements with a covering note, queries to a client about an unidentified transaction: this is real writing, in full sentences, with a tone that matters. It is also the writing most likely to be put off. Chasing money is uncomfortable, so the email sits in the mental queue for days. Speaking it out loud lowers the activation energy enough that it actually gets sent.
What you should keep typing
An honest guide has to include this section, because the fastest way to lose confidence in dictation is to use it where it does not belong. In accounting software the boundary is unusually clear.
Speak the sentences. Type everything else. Everything else, in Xero, means:
- Amounts. Never dictate a figure into a currency field. The risk-to-reward ratio is terrible: you save two seconds and expose yourself to a transposition you might not catch.
- Account codes and tracking categories. These are selections from a list, not prose. Use the keyboard and the dropdown, which will autocomplete faster than you can speak anyway.
- Dates, invoice numbers, reference numbers and tax numbers. Structured identifiers with rigid formats. Type them.
- Tax rates and VAT/GST treatment. Dropdown selections with legal consequences. Not a voice job.
- Email addresses and bank account details. Character-precise strings where a single wrong letter is a silent failure. See our guide on dictating email addresses and URLs for the cases where it is workable and the cases where it is not.
Read that list back and notice what it has in common: every item is a field where a small error produces a wrong number rather than an awkward sentence. Prose errors are visible and cheap to fix. Numeric errors are invisible and expensive. Keep voice on the side of the line where mistakes announce themselves.
Making it accurate on your client list
The first week of dictating into accounting software is usually a mixed experience, and the reason is almost always vocabulary rather than the transcription itself. General speech recognition handles ordinary English extremely well. It has no idea that your client is called Ardleigh Brookfield Holdings, that your practice abbreviates "work in progress" to WIP in every memo, or that the recurring supplier is a five-letter acronym.
Smart Vocabulary in Voice Keyboard Pro is a personal dictionary with replacement rules, and this is precisely the problem it exists for. Add the twenty or thirty proper nouns you say every day: client entity names, the directors you deal with, your recurring suppliers, and the internal shorthand your practice uses. Once those are in, the residual error rate on accounting dictation drops dramatically, because the errors you were getting were nearly all names.
If you want the deeper version of this, our post on building a custom dictation vocabulary covers how to choose which terms are actually worth adding rather than dumping your whole contact list in.
Two more accuracy habits that cost nothing:
- Speak in whole phrases, not word by word. Transcription uses surrounding context to disambiguate. Isolated words remove the context and accuracy falls.
- Say your punctuation. "Comma", "period", "new paragraph". It feels unnatural for a day and then it is invisible. Our guide to dictation punctuation on Mac has the full command reference.
A month-end workflow that uses voice properly
The best way to think about this is not "dictate instead of typing" but "batch the talking". Writing and data entry are different modes, and switching between them is what makes the work feel slow.
- Do the numbers first, silently. Work the bank feed, code the transactions, match what matches. Hands on keyboard, no voice at all. Leave the fields that need explanation empty or flagged.
- Do a second pass for prose. Go back through the flagged items and talk your way down the list. Memo, memo, reconciliation note, memo. You are in one mode now, and each note takes seconds.
- Then the client-facing writing. Invoice descriptions, the covering email, the summary of anything unusual you found. This is the most valuable writing you do all month and the part that most often gets rushed because you ran out of energy in step one.
- Capture on the phone as things happen. A note spoken into the Xero mobile app right after a client call is worth five notes reconstructed at month end.
The reason this ordering works is that dictation is fast at composition and slow at correction. Batching the composition into a run means you are never stopping mid-flow to fix a coded transaction, and the sentences come out better for it.
Fixing what comes out wrong
Nobody dictates a perfect paragraph every time, and the recovery step is where a lot of people give up on voice. On iPhone, Voice Edit handles this without touching the keyboard: select the text, speak the change you want, and it applies. "Make that more formal" or "change the second sentence to say the payment is now thirty days overdue" is faster than tapping a cursor into place on a phone screen.
On Mac the honest answer is that you have a keyboard right there. Dictate the paragraph, then fix the one word with your hands. Chasing a pure-voice workflow on a machine with a keyboard attached is a purity exercise, not a productivity one.
There is one tone trap specific to accounting work. Chasing an overdue invoice while mildly irritated produces a spoken paragraph that reads far sharper on the page than it sounded in your head. Speech carries tone; text does not. Dictate the chase email, then read it once before sending, and soften it if it landed harder than you meant. This is the same discipline that applies when dictating support replies, and it is worth building early.
Xero versus QuickBooks for dictation
If you run clients on both platforms, the practical difference is small. Both are browser-based, both accept system-wide dictation into any text field, and the speak-the-prose-type-the-numbers rule is identical. The field names differ, and Xero's reconciliation screen puts more short free-text opportunities in front of you than the QuickBooks equivalent does, which if anything makes voice slightly more useful on the Xero side. Our companion guide on dictating in QuickBooks covers the same ground for the other platform.
What matters more than the platform is that the same tool works in both, plus your email client, plus your practice management system, plus the browser tab where you are writing the year-end letter. A tool that only works inside one accounting package sits idle most of your day.
Speed, honestly
Most adults type at around 40 words per minute; professional typists reach 80 to 100. Comfortable speech runs at 130 to 150 words per minute. That gap is real, but it is not the whole story, and anyone claiming a clean three-times speedup on accounting work is overselling it.
The realistic gain is uneven. On a currency field, voice saves you nothing and costs you risk. On a two-sentence memo, it is meaningfully faster. On a paragraph of client explanation or a chase email, it is several times faster and, more importantly, it gets written at all instead of being deferred. Averaged across a bookkeeping day, the aggregate saving is solid, but the bigger effect is qualitative: your memos become explanations, your invoice lines become descriptions, and your contact notes stop being empty.
Privacy and client data
Bookkeepers hold other people's financial information, which makes this a fair question to ask before installing anything. Voice Keyboard Pro's server stores only operational pings. No audio and no transcript content is retained on our servers. The text you dictate goes to your cursor and stays in your systems.
That said, if you work under a client engagement letter, a practice policy or a jurisdiction with specific rules about processing client data, check them before you make dictation part of your standard workflow. We can describe how our product handles data; we cannot rule on your professional obligations, and you should not treat a blog post as if we could.
Frequently asked questions
Does Xero have built-in dictation?
No. Xero has no microphone button and no voice entry mode in either the web app or the mobile apps. Dictation comes from your operating system or a system-wide voice tool typing into Xero's ordinary text fields.
Does it work in the Xero mobile app as well as the browser?
Yes. On iPhone and iPad, a third-party keyboard with a mic button works inside the Xero app the same way it works anywhere else, because iOS keyboards are system-level. Tap the field, switch keyboards with the globe key, tap the mic.
Which browser should I use?
Any of them. System-wide dictation inserts text at the cursor, so it is indifferent to whether Xero is open in Chrome, Safari, Firefox or Edge. If dictation works elsewhere on your Mac but not in one specific browser, the usual cause is a keyboard shortcut collision rather than a browser incompatibility.
Can I dictate amounts if I am careful?
You can, and we would still advise against it. The saving is two seconds and the failure mode is a wrong number in a ledger that nobody notices. Type figures.
Will it get my clients' company names right?
Not on day one, and no general transcription engine will. Add them to Smart Vocabulary as replacement rules and they will be correct from then on. Proper nouns are the one error class you have to solve deliberately rather than wait out.
Where to start
Do not convert your whole workflow at once. Pick one field: invoice line descriptions. Dictate every description on your next invoice run and compare how long it took and how much better the descriptions read. That single change is usually enough to make the case, because it is the field where terseness costs you real money in query emails and slow payment.
If it sticks, add transaction memos next, then contact notes, then the emails. By the time you have those four, you have covered nearly all the prose in a Xero-based practice, and the numbers, which are the actual job, stay exactly where they belong: under your fingers.
Voice Keyboard Pro has a free tier with daily limits, so you can run that invoice-description test without paying for anything. Pro is $4.99 a month or $34.99 a year if you decide the talking is worth keeping.