Proof of Patrol App for Clients: Automatic Patrol Reports Instead of Paper Logs

A proof of patrol app replaces the paper patrol log with a record the guard's phone creates on its own: every checkpoint visit is stored as a completed task with the time, the GPS position and a photo checked by AI, and the site history for any shift or month becomes the report the client receives. The problem it solves is not that guards refuse to fill in the log. It is that a paper log proves nothing. A line that says "02:00 — round done — signature" was written by the person whose work it is supposed to confirm, at a time nobody can verify, and the client knows it. When a door is found open in the morning, the logbook turns into a piece of paper both sides argue about. If your rounds are already recorded with GPS check-ins and AI photo proof at each point, this article is about the next step: turning that record into a report a client accepts without a phone call, and doing it without anyone typing a report at the end of the shift.

Why the paper patrol log fails the client

A patrol logbook exists for one reader: the client who pays for the site to be covered. It is supposed to answer two questions — were the rounds done, and what was found. On paper it answers neither in a way the client can check.

The entries are written by the guard, often at the end of the shift from memory, and they look the same whether the round happened or not. The times are round numbers because the guard writes "02:00", not "02:07". Findings are compressed to "all in order" because writing more takes time and invites questions. And the book lives in the guard room, so the client only sees it during a site visit, or as a photographed page sent by the site manager once a month.

The result is a report that costs the firm effort every month and still cannot settle a dispute. The client's real question, "prove that the 2 a.m. round was done on the 14th", is answered with a signature.

What a proof of patrol app records instead

The change is not a nicer form. It is that the record is created at the checkpoint, by the phone, at the moment the guard completes the point — not written afterwards.

One entry per checkpoint, created at the point

Each checkpoint on the route is a recurring task in the team calendar with a due time and a location attached. When the guard completes it, the app stores the time and the device position. The task either exists as done at a verified position, or it becomes overdue when the due time passes.

Photo, position and time in one record

The task also carries a reference photo taken when the route was set up and a one-line description of the correct state — "gate chained and locked", "fire exit clear". The guard takes one photo, AI compares it with the reference and returns a score. The photo is stored with the position and the time, so a single checkpoint entry shows where the guard stood, when, and what the point looked like.

Exceptions logged when they happen

A photo that does not match the reference flags the task and sends a notification to the supervisor the same minute, with the photo attached. The guard adds a comment on the task — "padlock missing, called the site manager" — and that comment sits in the same history as the photo. The report later shows the exception at the time it occurred, not as a sentence written from memory at shift end.

Turning the shift history into the client report

Once every checkpoint occurrence is a task with a state, a time, a position and a photo, the report stops being a document someone writes. It is the history of the site filtered by period.

What the client actually wants to see

Clients rarely read a full round-by-round list. What they check is simple: every scheduled round exists and was completed on time; skipped or late points are visible rather than hidden; the exceptions have photos and a note about what was done; and the shift schedule shows who was on duty when each of those things happened. A history where done, overdue and flagged tasks are shown as such gives them all four without interpretation.

Weekly and monthly reports without typing

For a weekly or monthly report the supervisor opens the site history for the period and exports it. Each checkpoint occurrence is there with its state, time, position and photo, and the shift schedule for the same days sits next to it. For a dispute, the same history opened for one night answers the question in minutes: a real round is a chain of done tasks a few minutes apart at the right positions with photos that passed.

Setting up automatic patrol reports for a site

The steps below assume the route already exists; if it does not, set that up first.

• Build the route as recurring checkpoint tasks with a location and a reference photo at each point, as described in the guide to QR code patrol checkpoints and setting up a guard route. The report is only as good as this setup: a location attached from the office produces false flags the client will ask about.

• Write the description of each point in the words the client uses. "Warehouse door 3 closed and sealed" reads well in a report; "point 7 OK" does not.

• Agree with the client which exceptions they want to hear about immediately and which can wait for the report. Put the person who takes the immediate ones into the shift schedule so the notification has an owner.

• Ask guards to add a comment on any flagged or unusual task before moving to the next point — one sentence, while it is fresh. This replaces the "findings" column of the paper log.

• Decide the reporting period with the client — weekly for a new site, monthly for a stable one — and export the history on the same day each period.

• Before the first report, walk one round yourself and read the history from the client's side. If you cannot tell where the guard stood and what the point looked like, fix the setup before the client sees it.

Common mistakes when replacing the paper log

Keeping both. If the paper book stays in the guard room "just in case", guards fill it in instead of completing tasks, and the app history has gaps the client will notice. Remove the book on the day the route goes live.

Not reviewing flags before the report goes out. A flagged photo with no comment looks like a problem nobody handled. Ten minutes on the morning after each shift, in the first two weeks, keeps the report clean and trains the thresholds.

Assigning tasks to a named guard instead of the shift. When a replacement covers Tuesday night, the route must follow the shift, or the report shows a night with no rounds while a guard was in fact on site.

Treating the record as a stick. Guards accept a history that protects them from an accusation and resent one that exists only to catch them. Show the shift the first report and point out what it proves for them.

Put together, this is what a proof of patrol app looks like for a small security firm with a handful of ordinary sites: a printed label at each point, the phone the guard already carries, a checkpoint record with time, position and an AI-checked photo, and a site history that is exported for the period and sent to the client instead of a photographed page from the logbook. CosaNostra is built for exactly this: patrol routes as recurring tasks with locations and reference photos, notifications on overdue or flagged points, comments on the task where the exception happened, and a site history you can 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 is a proof of patrol app?

It is an app that records each checkpoint visit from the guard's phone as a completed task with the time, the GPS position and a photo checked by AI against a reference image. The history of those tasks for a shift, a week or a month is the patrol report — nothing is written up afterwards and nothing is installed on the site apart from printed labels.

What does the client receive instead of the paper patrol log?

The site history for the agreed period: every scheduled checkpoint occurrence with its state — done, overdue or flagged — the time, the position and the photo, next to the shift schedule showing who was on duty. It is exported for the period and sent to the client, so the same file answers a routine monthly review and a dispute about one night.

How are incidents and findings recorded without a "findings" column?

A photo that does not match the reference flags the task and notifies the supervisor immediately. The guard adds a comment on that task — one sentence about what was found and what was done — and the comment is stored with the photo and the time, so the report shows the exception at the moment it happened.

Can a guard still fake a round in the app?

Each entry is created only when the task is completed at a verified position with a timed photo, and every occurrence has a due time after which it becomes overdue. A round completed from one spot within a minute, a photo that does not match or a chain of overdue tasks is visible in the history, which is exactly what a signed line in a logbook could never show.