Dictate the order, review the draft, buy the label

The bench moment voice is for: hands full, screen far away

Every label tool assumes you are sitting down. The reality of a packing session is the opposite: one hand is holding a box flap shut, the other is stretching tape, the phone is propped against a roll of bubble wrap a metre away, and the order you are packing is a name you can see on a slip but would have to stop everything to type. Voice exists for exactly that posture. You say the order out loud — who it goes to, where, what is in the box, roughly what it weighs — and keep taping while a draft label assembles itself. The dictation is the capture step, nothing more. It does not price anything, it does not print anything, and it certainly does not spend anything.

This matters because the alternative is not typing carefully at a desk. The alternative, in practice, is a scribbled note that gets keyed in later that evening, from memory, in a batch, which is precisely the workflow that produces transposed digits and missing apartment lines. Speaking the order at the moment you are holding the parcel keeps the order and the object together. The wider intake story — pasting an order or screenshotting a DM thread and letting Billy draft from it — is covered end to end on the AI shipping labels page; this page is about the one intake mode that works when neither of your hands is free.

What a spoken order needs to contain

A usable dictation carries four things, and you can say them in any order and in plain speech. The recipient's name. The full destination — street line with the unit if there is one, city, state, and postcode. What the parcel is, in a phrase, because the contents description follows the shipment through customs forms and claims later. And a weight, which can be honest approximation: a figure read off the bench scale beats a guess, but a guess beats a blank, because the draft shows you the number and you can correct it against the scale before you buy.

You do not need to speak like a form. Saying it the way you would tell a colleague — the whole thing in one breath, hesitations included — is fine, because the parsing step is doing the structuring, not you. What the draft does with an incomplete dictation is the useful part: a missing piece stays visibly missing instead of being silently invented. An order with no weight shows an empty weight field waiting for a number, not a fabricated one. The draft is a checklist of what you actually said, which is exactly what you want to read back before money is involved.

Where mishears get caught: draft review and address validation before payment

Dictation has failure modes, and pretending otherwise would be selling you a mis-shipped parcel. Fourth Street becomes 4th Street, apartment B is heard as apartment D, a spoken postcode loses a digit to a tape gun screeching at the wrong moment. The design answer is not a better microphone — it is that nothing you speak is final. Everything lands in a draft first, printed back on screen field by field, and the checking step deliberately stays visual because errors that are inaudible are usually obvious in text.

Then the second net: before payment, the destination is validated against real address data, with a residential or commercial classification badge shown on the result. An address that does not resolve — a street number that does not exist on that street, a unit the building does not have — is flagged while the label is still unbought, which is the only moment a mistake is free. The mishear that survives both the draft read-back and validation is rare, and even then the label is usually voidable for a wallet credit rather than a write-off. The order of operations is the whole safety story: speak, read, validate, and only then pay.

Voice, paste, or screenshot: matching intake to the situation

Voice is one of three ways an order gets into a draft, and the right one is situational rather than a matter of taste. Paste wins when the order is already text in front of you and your hands are on a keyboard — a desk moment, not a bench moment. A screenshot wins when the order lives inside a DM thread and scrolling out the address would mean retyping it anyway. Voice wins when the parcel is in your hands: mid-pack, mid-tape, phone out of reach, or simply when the queue is long enough that walking to a keyboard between every box is the slowest part of the job.

The reason the three modes are interchangeable is that they converge immediately. Whichever way the order arrives, it becomes the same draft, gets the same address validation, and is quoted across the same four carriers as one all-in price each — the same comparison that makes a cheap label a choice rather than a default. And because Billy is included on every plan with no monthly minimum to print, dictating your two parcels a week costs exactly what the two labels cost. The intake mode is never the thing you pay for; it is just the door the order walked in through.