You step off the scale in the morning, still half-awake, water still on your skin, and you want to log the number before you forget it. Your phone is on the counter — or in another room. You're not going to dry your hands, unlock it, find the app, wait for it to load and tap your way to the right screen for a five-second task. So most mornings, you just don't.
That exact moment is what I built voice for. "Hey Siri, log my weight, eighty-one point four." That's it. No app, no unlock, no typing. The number is recorded before you've reached for the towel. Spend the last few weeks with me and you'd have watched me wire Gym Companion into Siri, Shortcuts and the Lock Screen so a handful of things happen without the app ever coming to the front.
Here's what surprised me: the interesting question wasn't how to add voice. It was where it actually earns its place — and, just as importantly, where it doesn't.
There are really only a couple of moments where talking to your phone beats touching it, and they share a shape. The task is small, you do it often, and your hands are busy with something else. The morning weigh-in is the cleanest example there is. A quick meal log is another — when you're mid-something and stopping to tap would break your flow, saying it out loud keeps you moving. Those are the moments I went after.
And then there's the part I think most apps would quietly skip over. I could have pushed voice into the workout itself — "log that set," muttered at my watch between reps. I built the plumbing; I could have done it. I didn't, because I can't honestly picture myself doing it. A loud gym, out of breath, surrounded by people — that is the last place I want to start talking to a device. Voice there isn't convenience.
It's friction wearing a convenience costume.
A feature that's perfect in one context can be faintly ridiculous in another. "Fully voice-enabled" would have looked good on a list of features. It also would have been a small lie to the people who'd actually try it at the gym and wonder why I thought that was a good idea.
I'll be straight about something else, too: voice is finicky to get right. Getting Siri to reliably hear the phrases, recognise them in the language you actually speak to it in, and stay registered after you restart your phone has taken more rounds than I'd like to admit — and I'm still sanding it down while the app is in beta. None of that is glamorous. But it's exactly the unglamorous work that decides whether a feature feels like magic or feels broken.
The job was never "add voice everywhere." It was to find the two or three moments where speaking genuinely beats tapping, make those feel effortless, and leave the rest alone. Restraint is a feature too. A good tool fits the moment you're in — and knows when to stay quiet.
More in two to three weeks.