Face Recognition Attendance · UAE

Face Recognition Attendance — From the Phone, Not a Kiosk

The biometric power of face recognition without the queue at a shared kiosk. Each staff member uses their own phone; face matching runs server-side; anti-spoofing built in; PDPL-compliant by default.

No kiosks · PDPL-compliant · Anti-spoofing · Human-in-the-loop · Free up to 25 staff

Face recognition was the right idea. The kiosk was the wrong delivery.

Face recognition is the correct biometric for attendance — identity-of-person rather than identity-of-finger, no shared touch surface, no hardware wear. The misstep most UAE deployments made was attaching the camera to a fixed kiosk at a single gate. That recreates the queue, recreates the hardware overhead, and adds a single point of failure on the day you don't want one. Aiya runs the same face recognition server-side, with each staff member capturing the selfie on their own phone. The biometric stays; the kiosk goes.

Server-side face matching

Live selfie at check-in is matched against the enrolled photo inside our own infrastructure. Not sent to third-party AI providers.

Paired with GPS verification

The face match confirms the person; GPS confirms the location. Both are recorded per check-in.

Anti-spoofing built in

Liveness checks reject printed-photo and screen-replay attempts. Impossible-GPS-speed and device-change signals are flagged.

Anomaly detection

The model learns each operation's rhythm; sudden swings in check-in time, location, or device surface for human review.

Human-in-the-loop

AI surfaces low-confidence matches for human confirmation. No black-box decision rejects a worker without a person looking.

Encrypted at rest + in transit

Industry-standard encryption. Role-based access. Biometric data never leaves our infrastructure for third-party processing.

Configurable retention

Biometric records auto-delete after a window your organisation sets. Audit history is kept separately, without the biometric image.

PDPL by design

Explicit consent on first use, encrypted storage, data-subject rights through the staff app. Aligned with UAE Federal Decree-Law No. 45 of 2021.

The face recognition flow, end to end

01

Enrol once

Staff member photographed in good light during onboarding. The encrypted face geometry is stored; consent is captured and timestamped.

02

Check in from any zone

Open the app at the work location, take a live selfie, confirm. Face + GPS verification runs in seconds.

03

Audit recorded

Match confidence, location, device, and time are recorded against the staff record with role-based access.

04

Anomalies surface

Out-of-zone, low-confidence, or unusual events surface for human review — never silently accepted or silently rejected.

05

Records auto-delete

Biometric data auto-deletes after the configured retention window. Audit metadata is retained.

Built around UAE compliance, not adapted to it

UAE PDPL biometric handling

Facial geometry is treated as biometric data: explicit consent on first use, encrypted at rest, processed in a UAE-aligned regional data centre, configurable retention.

No third-party AI providers see biometric data

Face matching runs inside our own infrastructure. Staff photos are never sent to external AI vendors.

Anti-spoofing

Liveness checks reject printed-photo and screen-replay attempts. Combined with GPS + device fingerprint, the system is harder to fool than a fingerprint reader.

Human-in-the-loop on edge cases

Low-confidence matches surface for human confirmation. No automated rejection of a worker without a manager looking.

Frequently asked questions

Is face recognition the same as facial scanning at a kiosk?
The biometric is the same; the hardware delivery is different. Aiya does server-side face matching with the selfie captured on the staff member's own phone. No kiosk at the gate.
How accurate is the match?
Approximately 98% in good conditions. Match quality is shown to the manager on every check-in so borderline cases are confirmed by a human rather than silently rejected.
Where is the biometric data stored?
In a UAE-aligned regional data centre, encrypted at rest. Never sent to third-party AI providers. Retention is configurable per your PDPL policy.
What about printed-photo spoofing?
Liveness checks reject obvious replay attempts. Combined with GPS zone check, device-binding signals, and anomaly detection, the system is harder to fool than a fingerprint reader.
Is there a free tier?
Yes — free up to 25 staff with the full face recognition + GPS feature set. Growth: AED 8/staff/month from staff 26+. Enterprise: custom (adds payroll + GPSSA + WPS file).

See face recognition without the kiosk

20-minute demo. We will enrol a face on a phone, run a live check-in, and walk through the PDPL retention flow.