Buddy-punching is a national sport in offices where attendance is just a signature. When building Dexova's attendance, the goal was not to surveil employees like prisoners, but to make presence data trustworthy enough to compute payroll without argument. That is a delicate balance.
Validate at the entry point, don't clean up later
Check-in happens through a PWA: the app reads location, finds the nearest office, and refuses check-ins outside the radius (geofence). Plus a mandatory selfie at check-in time — not a profile photo, but real-time proof of presence.
The principle is simple: cleaning dirty attendance data at month-end is far more expensive than refusing questionable data at the very first second.
Real WFH needs a two-address model
Naive geofencing breaks the moment WFH or field visits exist: employees are legitimately away from the office. The answer is a two-address model — a fixed assigned office and a temporary work location — with an approval flow.
An employee outside the office can choose: check in as a field visit (temporary, for a set period) or request a permanent reassignment that needs HR approval. The policy lives in data, not hard-coded per company.
Corrections always go through approval
Wrong check-in times will happen — bad signal, forgetfulness, a laggy app. What must not happen: an employee or admin editing the attendance number directly in the database.
Every attendance correction is a request that passes through approval, leaving a trail of who changed what and why. That is what lets the monthly recap be locked with confidence before it feeds payroll.
Summary
Trustworthy attendance is not about the fanciest technology, but about refusing bad data at the source, accommodating how people actually work (WFH), and ensuring every change leaves a trail. Clean attendance data is the foundation of undisputed payroll.