Dictation is not the same as voice invoicing
The phrase "invoice by voice" covers two quite different things, and the difference decides whether it saves you any time at all. The first is dictation: your phone converts speech to text and you paste that text into a box. You still have to structure it, price it, and put it in the right fields — you’ve changed how the characters get typed, not how much work there is.
The second is what’s actually useful: speech that becomes a structured document. You describe the job in normal language and the result is separate line items, each with a quantity and a price, tax applied where it belongs, attached to a customer. That’s a different thing, and it’s the one worth looking for.
What to actually say
The single biggest improvement people make is describing the job the way they would to the customer standing next to them, rather than trying to speak in form fields. "Replaced the kitchen faucet, about an hour of labour, one new mixer tap, plus the call-out" works. Reciting field names does not — it’s slower, it feels unnatural, and it produces worse results.
Say quantities out loud where they matter ("two hours", "forty square feet", "six doors"), because a quantity you don’t say is a quantity something has to guess. And name the extras explicitly as you go, since the whole point is to catch the things that otherwise get remembered at the end of the month, or not at all.
Pricing is where voice either works or falls apart
Speech turning into line items is only half the job. Those line items then need prices, and there are only two places a price can come from: a rate list you set up in advance, or a guess.
This is why a price book matters more than the voice feature itself. With your real rates stored — labour by job type, common materials, call-out and travel fees — "replaced the kitchen faucet, one hour" resolves to your actual numbers every time, and two similar jobs six weeks apart get quoted consistently. Without one, voice capture just produces a nicely formatted document full of numbers you still had to invent.
Review before you send — always
Any speech system will occasionally mishear, particularly on a windy site, in a van with the engine running, or around trade-specific words and brand names. That’s manageable as long as reviewing the draft is a habit rather than an afterthought.
Treat the output as a draft that is ninety percent right, not as a finished document. Scanning five line items takes a few seconds and catches the one that says eleven when you said seven. The time saved by voice is in not typing; it was never in not checking.
When it genuinely beats typing
Voice wins hardest in three situations. On-site, when your hands are dirty and the alternative is doing it tonight. Between jobs, when you have ninety seconds at the kerb and a form would need ten minutes. And on the jobs with lots of small parts, where speaking a list is much faster than tapping out eight separate lines.
It wins least when you’re already at a desk with a keyboard and a complicated job in front of you. If you’re sitting down to build a detailed multi-stage quote, typing is fine, and pretending otherwise would be silly. The argument for voice is about the paperwork that never happens because it requires sitting down at all.
The real test
The measure that matters isn’t whether the transcription was perfect. It’s whether the quote went out the same day. A quote sent from the driveway beats a more elegant one sent two evenings later, because the customer has usually decided by then — often based on whoever answered first.
Voiczer is built around that specific gap: describe the job out loud, have it priced against your own rate list, check the draft, and send it before you’ve left. The voice part is the mechanism; getting the document out while you’re still standing there is the point.
See it running on your own price book
Voiczer prices jobs from a spoken description against your own rates, on-site.
Join the waitlist