CAM-522
Stop scrubbing through minutes: search for what happened
An incident in a branch is always told as a sentence: “somebody came in behind an employee through the staff door,” “a customer fell beside the ATM,” “that box was taken off the counter at ten at night.” The recording exists: the branch’s four cameras are there, and last night’s too. What doesn’t exist is the whole afternoon it takes to watch them. At this desk the sentence is typed and IRIS returns the moments when it happened, each with its camera, its date and its time. It doesn’t search for people: it searches for facts.
Turn on IRIS Vision
AI-generated image
What IRIS is
IRIS keeps an eye on the branch for you.
IRIS is an artificial intelligence that watches the cameras you already have at the branch and understands what's happening at the lines, ATMs, and teller windows. When a line gets long, a window is left unattended, or something you've asked it to watch happens, it alerts you — and whatever can be counted, like people waiting, it counts in numbers. All day long, with no one needing to stare at a screen.
You tell it what you need to watch, and IRIS does it.
Instead of scrubbing through a whole shift of video, here you just tell IRIS what happened in one sentence — someone slipped in behind an employee, a customer fell by the ATM — and it hands back the exact moments, each tagged with its camera, date and time. Staff no longer burn an entire afternoon staring at footage to find one fact. You use those minutes to decide whether to file a report or close it out, and you confirm it in a few minutes flat.
Works with the cameras you already have. Nothing to replace.
A person replies, not a form.
- Your footage never leaves your premises.
- It doesn't identify people.
- Certifications →
What IRIS understands
What it recognizes in this scene
- video wall of the room
- operator at the left desk
- operator at the right desk
- monitors at the left desk
- monitors at the right desk
- operations telephone
- window of the room
What this camera measures
An alert says something’s happening. The reading says how much.
- 20:00 – 24:00Stretch described
- 4Cameras in that stretch
- 16 hFootage to be reviewed
- 4Moments returned
- 21:38:52Clip opened at
The figures on this screen are an example from the illustration, not any client’s measurement.
The rule
You decide once. IRIS applies it always.
If
somebody describes in one sentence a fact and the stretch of time when it may have happened
Then
- return the moments when that happened, with camera, date and time
- open each one in the video, at the exact second
- export the clip and record who asked for it and when
The instruction that set it up«Show me the moments when somebody went through the staff door behind another person, in this branch, between eight in the evening and midnight yesterday.»
What changes
What changes for whoever answers for the branch network
The expensive part isn’t recording: the expensive part is still having to look. A customer complaint, an insurance report or a court request are settled today by one person in front of four screens for a whole afternoon, and that’s paid in the hours of somebody who is needed elsewhere. When what you look for is a fact and not a minute, that afternoon shrinks to the time it takes to type the sentence and check four results. And what comes out is a clip with its camera, its date and its time, exportable as it is, and with a record of who asked for it and when.
What people usually ask
Does IRIS search for a particular person?
No. Here IRIS searches for facts and moments, not people: one person going through a door behind another, a fall beside a ATM, an object taken off a counter at a certain time. What you write is what happened and when, not who. IRIS does not recognize faces, does not compare them with anything and keeps nobody’s identity. Who it was, if that needs to be known, is worked out by a person watching the video, exactly as they would today — except they already know which minute to watch.
And what if I don’t remember the exact time?
That’s the normal case, and it’s exactly the problem being solved. Nobody remembers a time: people remember what happened. A rough window — “yesterday afternoon,” “Monday at closing” — plus the description of the fact is enough; you narrow it down with what comes back. Searching by time forces you to know the time, which is why today it ends up as six hours of four cameras. Searching by fact doesn’t.
What can it not find?
It can only find what some camera saw and what’s still recorded. If the spot had no camera, if the frame didn’t reach that far, or if the recording has already been deleted under the retention period the institution itself sets, there’s nothing to find — and IRIS says so instead of inventing it. Nor does it decide: it proposes moments and a person confirms which one is right, and that confirmation is signed. A system that decided on its own would be no use for taking anything to court.
Other cameras
- CAM-510 · The line tells you how many desks you need at that hour
- CAM-511 · Five people waiting and a desk set up with nobody at it
- CAM-512 · Two people through the door on a single opening
- CAM-513 · A restricted door standing open and someone waiting outside
- CAM-514 · An operation that takes two people, and there’s one in the room
- CAM-515 · Someone has been in the ATM lobby for twenty-two minutes
Simulation · how IRIS works
You set the rule once. IRIS applies it every time.
You will see a summary of the interface, played step by step and hands-free. Each round starts with another case. The complete tool doesn’t fit in a demo.
- The rule gets built
- It goes to every camera
- Published, it watches alone
- And then it fires
Next step
Who sees the sales that never reach the register?
10 cases: 4 raise an alert when something happens and 6 only measure. You see each image as it is, and then with what IRIS understands on top.
Who is watching when the load swings over somebody?
11 cases: 9 raise an alert when something happens and 2 only measure. You see each image as it is, and then with what IRIS understands on top.
And the best part
Video is late by definition:you watch it once the thing has already happened.
IRIS tells you while it’s happening, with the camera, the time and what it saw. Arriving in time changes the end of almost any story, and no new cameras are needed for that.
Tell us your problem and we’ll say whether IRIS understands it, or not yet.
A person answers, the same working day.