Proof of Visit: How to Show a Field Employee Was Actually On Site
A customer queries an invoice for a Tuesday callout. Your engineer says he was on site from nine until half past eleven. The timesheet says the same thing, which is unsurprising, because he wrote both. You believe him. That is not the problem. The problem is that the only thing you can send the customer is a screenshot of your own spreadsheet, and they know it.
Running this manually today?
See how Fieldproxy handles it — a walkthrough on your own workflow, not a canned demo.
Book a demoAlmost every record of a field day is written by the person it describes
That single fact explains most of what goes wrong with field records. Clock-ins get forgotten on the days that are busy enough to matter. Durations get rounded to the nearest half hour, always in the same direction. Travel time is estimated after the fact. And the site nobody reached does not appear in the record at all, because nobody writes down the visit they did not make.
None of that requires anyone to be dishonest. A record reconstructed on Friday afternoon about a Tuesday morning is simply a weaker kind of evidence than one created at the moment it describes, and everyone in the dispute understands the difference.
What actually counts as proof
A record that survives being challenged usually has five properties:
- A timestamp created at the moment of arrival, not typed in later
- A position fixed by the device itself, rather than a site picked from a dropdown
- A departure captured the same way as the arrival, so duration is derived rather than claimed
- Something continuous in between, so a five minute appearance cannot be presented as two hours
- Visibility of the visits that did not happen, not only the ones that did
Four ways teams capture it, and where each one breaks
| Method | What it actually proves | Where it breaks |
|---|---|---|
| Paper sheet or a WhatsApp message | That somebody wrote something down | Written from memory, often days later. There is no independent timestamp anywhere in it. |
| Photo taken at the site | That a photo was taken somewhere | Says little about when, nothing about how long, and one photo can be reused for several visits. |
| QR or NFC tag at the door | That a phone touched that specific tag | Captures one instant and nothing either side of it. The tag can also be photographed once and scanned from the car park. |
| Geofenced check-in app | That someone pressed a button inside a radius | Still a deliberate act, so it is forgotten on exactly the days you most need it. It can also be pressed while driving past. |
| Continuous location trail | Where the device actually was, through the day | Needs the app running and the battery to hold, and it requires an open conversation with the team about what is being recorded. |
The failure nobody plans for: the location is faked
Mock location apps are free, they do not require a rooted phone, and they are widely used precisely because they defeat attendance systems. To a check-in app, a spoofed coordinate is indistinguishable from a real one. The app was only ever designed to ask where the device says it is.
If your evidence rests on a single tap at a single moment, a mock location app defeats it completely and you will never know. A continuous trail is much harder to fake convincingly, because the whole day has to hang together, including plausible travel time between the stops.
Derive the visit instead of asking for it
The alternative to asking someone to record their day is to derive the record from the location trail itself. Arrival, departure, duration and travel all fall out of the readings, and no one has to remember anything. The useful side effect is coverage: because every stop is derived rather than declared, the sites nobody reached are just as visible as the ones they did.
One tool built specifically around that idea is VisitsTracker, which produces a visit for every stop from the underlying readings rather than from taps, flags devices reporting a faked position, and shows which planned sites went unvisited. It is worth a look if your current answer to a disputed invoice is a spreadsheet. You can put it on a single phone and run one ordinary week before deciding anything.
Have the privacy conversation first, not after
Continuous location tracking of employees is lawful in most places when it is limited to working hours, disclosed clearly, and tied to a genuine business purpose, but the specifics vary by country and by state, and some jurisdictions require documented consent. Tell the team what is recorded, when recording stops, and who can see it, before anything is switched on. A system the crew discovers by accident will be worked around, and a record the team does not trust is not much use in a dispute either.
Common questions
Does a QR code checkpoint prove someone stayed on site?
No. It proves a phone touched that tag at that moment. It says nothing about the hours either side, which is usually the part being disputed. Checkpoints work best as a supplement to a continuous record rather than as the record itself.
What is acceptable as evidence in a customer invoice dispute?
There is no universal standard, and most disputes are settled commercially rather than legally. In practice what settles them is a record created independently of the person being questioned, with arrival and departure captured the same way, that you can send over without editing it first.
Can employees see the same record their employer sees?
They should be able to, and in several jurisdictions they have a right to request it. It is also good practice regardless of the law, because a worker who can check their own day will correct an error early rather than dispute the whole system later.
Is continuous tracking legal?
Generally yes during working hours, with clear notice and a legitimate purpose, but the rules differ meaningfully between countries and between US states, and a few require explicit written consent. Check the position where your staff actually work before you roll anything out.
See what a tracked day actually looks like
Live positions, a visit for every stop with time in and time out, and the sites nobody reached. Run it on one phone for a week.
Visit VisitsTracker