See who's there
A face, a badge, or a quick QR scan tells the room who just walked in. This is the only step that involves any guessing, and we keep it right at the front door. Whichever method is used, everything that happens after is the same.
How it works
From a face at the door to a room that's ready.
The whole thing is five plain steps that always run in the same order. Either a person walks in, or a meeting gets booked on the room's calendar. From there it's the same path every time: see who's there, look them up, follow the written rules, set the room, and write down what it did.
Look up who they are
The room checks that person against the company directory and reads their department and role. Now it knows not just who is present, but what they're most likely there to do. Nothing here is invented; it's a straight lookup, and it gets written down.
Apply the written rules
A fixed list of rules turns that role into a setup: a presentation, a working session, a video call, an executive briefing, or a simple welcome for a guest the room doesn't know. The rules are written down in advance, so the same person gets the same setup every time. There is no model deciding on the fly.
Set the room
The room carries out the setup on its own, pointing the camera, choosing the right screen and source, and adjusting the lighting, then checks that the change actually took. No panel, no button, no one reaching for a remote.
Write it down
Every time, the room records who it recognized, which rule it followed, what it set, and whether the room confirmed the change. That record is why a business can let a room run itself and still answer, exactly, for what it did and why.
Why the rules matter
Same input, same result. Every single time.
Step three is the heart of it, so it's worth being clear about what "written rules" means and why it's the part that earns trust. A rule is just a plain if-this-then-that, decided ahead of time and written down: if a sales lead walks into this room, set it for a client video call. Because the rules are fixed, the room can only ever respond one way to the same situation. Run it a thousand times and you get the same answer a thousand times.
That predictability is the whole point. A chatbot-style assistant might phrase things a little differently each time, which is fine for conversation and a real problem for a room full of equipment. Here, the AI never gets to decide what the room does. It only recognizes a person, then hands off to rules that behave the same way today, tomorrow, and a year from now, and a record of every step is kept so anyone can read back exactly what happened. Nothing weird, and nothing you can't explain.
Two ways it gets triggered
Proactive before you arrive, reactive when you do.
Something has to set the five steps in motion, and there are two kinds of trigger. The room can reach ahead of you, or it can respond to you in the moment. Either way, once it's triggered it runs the very same rules and keeps the very same record.
It gets ready before anyone arrives
The room watches its own meeting calendar and reads what's coming: who's invited, the kind of meeting, and any saved preferences for the people on it. The moment a meeting is booked, moved, or cancelled, the room notices and starts getting itself ready, so it's set before the first person walks in. No one has to be standing there for it to act.
It responds to who it senses in the moment
When someone is actually there, the room responds to them: a face it recognizes, a badge or QR scan, occupancy and proximity sensors that say the space is in use. These in-the-moment signals are how the room reacts to a person who walked in without a thing on the calendar.
What matters is that the two kinds meet at the same place. Whether a calendar invite triggered it ahead of time or a person triggered it on arrival, the room runs the exact same five steps and follows the exact same rules. If a piece of gear happens to be offline, the decision and its record still stand, the room just notes that the change didn't take. And the calendar watch renews itself automatically, so the room never quietly stops paying attention. Proactive or reactive, the room behaves the same way, which is the point.
The engine underneath
The rules engine
Every step above is carried out and recorded by one engine that runs the rules and keeps the log. It is the reason the room behaves the same way every time, and the reason there is always a record to read back. The AI gets to recognize a person. The engine decides what happens next and writes it all down. The AI only recognizes you. Written rules decide and act. If you want the full picture of how that engine is built, the Resources docs lay it out in technical detail.