What we tell every client up front.

Before anyone signs anything, we say the quiet parts out loud. A signless, touchless interface is a real upgrade — and it has real limits. Here are the four we put on the table in the first conversation.

It would be easy to sell a touchless, signless interface as magic. It isn't. It's a careful, honest upgrade that works beautifully inside a clear set of boundaries — and the fastest way to lose a client's trust is to discover those boundaries after the install. So we name them first.

Signing isn't universal — so the alternatives stay

Not everyone signs, and not every interaction should require it. A signed PIN is a wonderful option; it is never the only option. Tactile buttons stay. Spoken alternatives stay. The touchscreen stays. Every interface we add is one more door into the same room, never a door we've quietly locked behind us. (This is principle one for a reason — the full six are here.)

Privacy matters — so signed input pairs with a second factor

A camera that reads your hand is more private than a microphone that broadcasts your voice across a lobby — but "signed in the air" is not the same as "secure." For anything that protects money, access, or identity, a signed PIN is one factor, paired with a second — a card, a token, a known device. We design the camera to recognize the body in front of it, never to store it: frames are read and discarded on-device, and nothing about a face or a hand leaves the room.

Recognition has limits — so we design for what's ready today

We build on static handshapes, digits, and a small confident gesture set — the parts of the field that are dependable right now. We do not promise continuous sentence translation or flawless reads in every lighting condition. When a sign is missed, the system fails the way a good interface should: it says so plainly, offers another way through, and never strands the person at the counter. We design for graceful degradation, not for a demo-day miracle.

The room matters — so lighting and placement are part of the build

A recognizer is only as good as the frame it's handed. Backlight from a window, a camera mounted too high, a glossy counter throwing glare — any of these will hurt a read no software can rescue. That's why lighting, camera placement, and sightlines are part of every install, scoped on the floor with you, not assumed away in a spec. We solder the boards and we walk the room.

The kindest interface is the truest one — and the truest one tells you its limits before you find them.

None of this is a reason not to build. It's the opposite: these are the four conversations that turn a flashy idea into something that still works on a rainy Tuesday a year from now. Say them out loud early, design around them honestly, and the result is an interface people actually trust.