A product demo fails in a different place from a talking-head video. The words are usually fine — you know the product, you could describe it in your sleep — and the take still dies, because the failure lives in your hands. The moment arrives where the sentence says now let me show you, and the thing you are showing is face down, or out of reach, or the cap is still on, or the cable is coiled on the far side of the desk. Show-and-tell formats — product demos, unboxings, tutorials — are built entirely out of these small handovers between speech and action, and every handover is a place where the take can quietly fall apart. Watch your own outtakes from the last shoot and you will find the pattern already there: it is almost never a fluffed line that sends you back to the top of the script. It is a fluffed move, performed by hands that were never given a script of their own.
The instinct is to wing those moments, and the instinct comes with a respectable-sounding excuse: the actions feel too obvious to plan. Of course you will pick the phone up when you reach the part about the phone. But demonstration is choreography whether you plan it or not — the object has to be in position before the sentence starts, facing the lens rather than facing you, arriving in frame as the words land rather than two seconds after them. Left unplanned, each of those alignments is a dice roll, and a demo contains a dozen of them, so the odds compound quietly against a clean take. Worse, an action error is invisible to the edit. A misspoken line can be re-said and trimmed; an object that never entered the frame was simply never filmed, and there is nothing in the footage to cut your way around. The only repair is another take, and then another, which is how a three-minute demo becomes an afternoon.
The fix is to plan the moves in the same document as the words — not in a separate shot list you cannot look at mid-take, but in the script itself, at the exact sentence where each move belongs. In Talk2Camera, a line written like [SHOW THE PRODUCT] becomes a stage-direction chip in the prompter: it scrolls up with the script, visibly apart from the spoken sentences, and it exists to be shown, acted on, and never spoken. Your choreography arrives in front of your eyes at precisely the moment it is needed, which no checklist taped to the tripod has ever managed. And because a chip does not read as prose, the reflex that speaks whatever the prompter shows leaves it alone — the classic failure of writing REMEMBER THE LID into a script and then warmly telling your audience to remember the lid is closed off before it can happen.
The plumbing underneath matters more in a demo than anywhere else, because demos are full of pauses. A chip cannot reach voice follow — the speech recognizer never waits for words you were never going to say — so while your hands are busy and your mouth is quiet, the scroll is not stalled on a line nobody will read out; it is simply waiting for your next spoken sentence, and it picks you up the moment you speak again. And a chip cannot reach the captions: the caption track stays clean, with no bracketed instructions leaking into the subtitles of the finished video for a sharp-eyed viewer to screenshot. You can salt a script with as many directions as the demo genuinely needs — lift here, turn it here, point here — and not one of them will ever be heard, waited for, or published. The prompter is the only place they exist, and you are the only person who will ever know how much choreography the effortless-looking take actually contained.
There is also no need to start the choreography from a blank page. Talk2Camera ships script templates — eight starting points, covering a product announcement, a tutorial, a weekly update, a two-presenter interview and more — and each one is a real spoken scaffold with slots to fill in, written in your own language rather than in placeholder-ese. A tutorial template already has the shape a tutorial needs: what the viewer will learn, the thing itself, the steps, the place where the steps usually go wrong. Your job shrinks to filling the slots with your product and dropping stage directions where your hands have work to do. The structure that takes most people a hundred demos to internalise arrives pre-installed, and your planning time goes to the ten seconds of the video that are genuinely yours and nobody else could script.
A few honest limits. A chip tells you when; it cannot teach you how, and a physical move that has never once been done smoothly — peeling a screen protector, opening clamshell packaging — will not become smooth because a prompter mentioned it. Do the fiddly action once, off camera, before you roll. Keep directions short and imperative; a chip you have to study mid-take is a chip that steals your eyes at the worst possible moment. And expect the first take to reveal a missing direction or two — the object you always forget, the step you always transpose — which is fine, because the script is now where such discoveries belong: add the chip, and that mistake is spent for good. The demo that used to take nine takes starts taking two, not because your hands got better overnight, but because they were finally handed the same script your mouth has had all along.