Security Incident Reporting From the Guard's Phone: What to Log and How
Security incident reporting from the guard's phone means the guard records what happened at the moment it happens — with a photo, the time, the position and a spoken description that the app turns into a structured card — instead of writing a report from memory at the end of the shift. The operational problem is familiar to any owner of a small security firm: at 02:40 a guard finds a warehouse door open, calls the supervisor, tapes it shut and moves on; at 07:00 the client asks what happened, and the only record is a line in the logbook written at 06:30 that says "door 3 open, secured". No photo, no exact time, no idea whether anything was taken. If your rounds are already recorded as an automatic patrol report from the guard's phone, incidents are the piece that still lives on paper — and this article is about moving it.
Why incident reports written at shift end are useless
An incident report has two readers: the supervisor who has to decide what to do now, and the client who will ask about it later. A report written at the end of the shift serves neither.
By 06:30 the guard has done four more rounds. The time becomes "around 2:30", the description shrinks to one sentence, and the details that matter — was the padlock cut or missing, were the lights on inside, which way the tracks in the snow went — are gone. The supervisor learns about the incident from a phone call and has nothing in front of him. The client gets a sentence and a signature, and a disputed claim turns into two stories with no evidence on either side.
The fix is not a longer form. Guards will not fill in a long form at 02:40 in the cold, and they should not have to. The fix is to capture three things at the point, in under a minute, and let the phone do the rest.
What to log in security incident reporting
Every incident record needs the same four elements, and none of them should require typing.
Photo first, before anything is touched
The photo is the evidence. It should be taken before the guard closes the door, moves the object or steps into the area — the state as found, not the state after the guard fixed it. Two or three photos are better than one: the wide shot showing where it is, the close-up showing what is wrong. If the guard's rounds already use AI photo checks at each checkpoint, the guard is used to this, and the incident photo simply goes on the same task.
Time and position, recorded by the phone
The time and the location must come from the device, not from the guard's memory. A record created at 02:41 at the position of warehouse door 3 is a fact; "around 2:30 at the warehouse" is a recollection. This is the same principle that makes security patrol tracking without RFID work: the phone already knows where and when, so the guard should never be asked to enter it.
What was found and what was done, spoken, not typed
The description is where reports fail, because typing on a phone with gloves on at night does not happen. The guard should be able to say it: "Warehouse door 3 open, padlock missing, no one inside, closed the door, called Mark." Voice AI turns that sentence into a structured card — a task with the site, the time, the person to notify and the photo request — and the guard confirms it with one tap. What the guard said stays as the description, in the guard's own words, which is what a client or a police officer will later want to read.
Who was told, and when
An incident record without an owner is a note nobody acts on. The card must name the person who was notified — the supervisor on that night's shift — and the notification must go out the moment the card is saved, not when someone reads the logbook in the morning. The supervisor then sees the photo, the position and the description on his phone and decides whether to drive out or wait for the day shift.
How to set up incident reporting from the guard's phone
The steps below assume the guards already carry a phone with the app used for their rounds. Nothing is installed on the site.
Decide who receives incident notifications on each shift and put that person into the shift schedule. If the supervisor changes on weekends, the notification must follow the schedule, not a name typed into a form.
Write a short rule for guards: photo first, then speak, then act. Three lines on the wall of the guard room are enough. The order matters — a photo taken after the door is closed proves nothing.
Give the guards three example sentences to say, in the form the app understands: what, where, what was done, who to tell. "Fire exit on the second floor propped open with a brick, removed it, photo attached, notify the supervisor." Guards copy patterns; give them good ones.
Agree with each client which incidents they want to hear about immediately and which can wait for the shift report — an open door at 3 a.m. and a broken lightbulb are not the same call.
Ask guards to add one comment on the incident task when it is closed: what the outcome was. "Site manager arrived 03:20, door locked, nothing missing." That closes the record.
Run one drill on each site in the first week: stage a fake incident, time how long the record takes, read it from the client's side. If it takes more than a minute or reads badly, fix the example sentences.
Mistakes that make incident records worthless
Letting the guard write the time. Any field the guard fills in by hand is a field a lawyer can question. Time and position come from the phone or they do not count.
Keeping a paper incident book as a backup. If both exist, the guard fills in the one that takes less effort, and the app history has gaps that the client will notice on the day it matters.
Reporting to a person instead of a shift. The supervisor who was on duty last Tuesday is on holiday this Tuesday. The notification must go to whoever is in the schedule tonight.
Making incident reporting a punishment. If a report leads to a lecture, guards stop reporting. A record with a photo and a timestamp protects the guard from the accusation that he did nothing; say that out loud at the next shift briefing.
Where the incident records live afterwards
Once each incident is a task with a photo, a time, a position, a spoken description and a closing comment, the history of a site becomes the incident log the client asks for. Open the site for a month and every incident is there in order, next to the rounds that were done and the shifts that were on duty, so a question like "how many times was the back gate found open in August?" is answered by looking, not by reading a paper book page by page. The same history exported for a period is what goes to the client together with the patrol report, and what goes to the insurer or the police when a single night has to be reconstructed.
That is what security incident reporting looks like for a small firm when it is done from the guard's phone: a photo taken as found, a time and a position the guard never typed, a description spoken in the guard's own words and turned into a card by AI, a notification that reaches the supervisor on that shift within the minute, and a closing comment that ends the record. CosaNostra is built around exactly these pieces — voice creation of tasks with a site, a time, an assignee and a photo request, photo attachments and comments on the task, notifications to the person on shift, and a site history you can open or export for any period, with shifts in the same calendar. See the full setup for security companies on the CosaNostra guard tour system page.
Frequently asked questions
What should a security incident report from the guard's phone contain?
Four things: a photo of the situation as found, taken before anything is touched; the time and position recorded by the phone rather than typed by the guard; a description of what was found and what was done, spoken and turned into a card by the app; and the name of the person who was notified. A closing comment with the outcome completes the record.
How does security incident reporting work without typing?
The guard takes the photo and says what happened in one sentence — what, where, what was done, who to tell. Voice AI converts that sentence into a task with the site, the time, the assignee and the photo attached, and the guard confirms it with one tap. The spoken words stay as the description, so nothing is retyped and nothing is written from memory at shift end.
Who gets notified when a guard logs an incident?
The person who is in the shift schedule for that night — the supervisor or the senior on duty — receives a notification with the photo, the position and the description the moment the card is saved. Because the notification follows the schedule and not a name in a form, a weekend replacement receives it without any change to the setup.
How does the client receive incident records?
Each incident is stored in the site history as a task with its photo, time, position, description and closing comment, next to the rounds and shifts of the same period. The history is opened or exported for the agreed period and sent together with the patrol report, and the same export for a single night answers a dispute or an insurer's request.
