Zaffre HRM
← All articles

How to Do Face Recognition Attendance on Mobile — Setup Guide

Zaffre HRM Team · May 30, 2026

Setting up face recognition attendance on mobile is much simpler than buying and installing biometric hardware. Once the company-level configuration is done (15 minutes), each employee enrols in 30 seconds. After that, attendance is just a phone glance. Here is the practical setup guide for Pakistani SMBs.

Phase 1 — Company-level configuration (15 minutes)

1. Enable face recognition module

In the admin console, enable face recognition attendance for the company.

2. Set liveness check requirement

Recommended: ON. This is what prevents photo / video spoofing. The slight extra second at check-in is worth the anti-fraud protection.

3. Set geo-fence policy (optional)

If you want attendance restricted to specific locations, define the geo-fences:

  • Office location — specify coordinates + radius (typically 100-500 metres)
  • Per-location for multi-site businesses
  • "Off" for distributed / remote teams where location does not matter

4. Set time windows

When can employees check in? Often 6am-10am for morning shift, etc. Outside the window, check-in is rejected (or marked late).

5. Set device binding (optional)

Restrict each employee's face check-in to their registered phone. Prevents someone else from using their phone for check-in.

6. Configure retention

How long face data is kept after employment ends — typically 90 days for dispute resolution, then deletion.

7. Update WFH / attendance policy

Make sure the employee handbook reflects the face attendance setup, consent, and rights. See: biometric legality in Pakistan.

Phase 2 — Per-employee enrolment (30 seconds each)

For each employee:

  1. Open the HRMS mobile app on their phone
  2. Go to attendance → face enrolment
  3. Hold the phone at face level, well-lit area
  4. App captures 3-5 face angles (front, slight left, slight right)
  5. Random micro-prompt (slight head movement) for liveness baseline
  6. System generates face embedding (numerical vector)
  7. Enrolment confirmed

Best practice: do enrolment in well-lit conditions. Avoid bright backlight (windows behind face). Glasses on if usually worn.

Phase 3 — Daily use (3 seconds per check-in)

  1. Employee opens HRMS mobile app at start of work
  2. Tap "Check In"
  3. App opens camera, prompts for face capture
  4. Employee looks at camera
  5. Liveness check (motion / depth) — auto-passes for a real face
  6. Face matched against enrolled embedding
  7. Attendance recorded; push confirmation

Edge cases to handle

Failed match

If face does not match (lighting issue, large appearance change), the system rejects. Employee can retry. After 3 fails, fallback to manager-approved manual entry.

Appearance changes

Beard grown, glasses changed, weight changed — re-enrol. Takes 30 seconds.

New phone

If device binding is enabled, employee registers their new phone via a one-time code from HR.

Lost phone

Web kiosk fallback at office. Face enrolment data is in cloud, not on the phone, so a lost phone is not a lost enrolment.

Bad lighting / no signal

App captures attendance offline, syncs when connection restores. Useful for field staff in patchy areas.

What employees should be told

  • "Face data is stored as embeddings, not photos. Not reversible to a photo."
  • "Data is encrypted and only accessible to HR + IT in audit role."
  • "You can re-enrol any time."
  • "On exit, your data is deleted after [X] days."
  • "You can ask to see your own enrolment data."

Common setup mistakes

  • Skipping liveness check — buddy punching becomes trivial again
  • Geo-fence too tight (50 metres) — false rejects when employee is just outside
  • Time windows too narrow — late employees cannot check in
  • No handbook update — employees feel ambushed
  • No re-enrolment process — employees stuck after appearance change

The Zaffre HRM mobile face attendance fit

Zaffre HRM mobile app ships with face recognition + liveness check + geo-fencing + device binding + offline capture. Per-employee enrolment in 30 seconds. See: face recognition explained.

Book a demo for live face attendance setup.