AI Attendance · UAE

AI Attendance — Face Verified, From a Phone

Face-verified, location-checked attendance from each staff member's own phone. No shared fingerprint readers, no kiosks, no morning queue. AI does the verification; humans confirm every consequential decision.

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

Biometric proof of presence — minus the queue at the door

Traditional biometric attendance means a shared scanner, a queue, recurring hardware costs, and a long list of failure modes — humidity, dust, broken readers, shared touch surfaces. The phone in each staff member's pocket already has a camera, GPS, and a connection. Aiya uses that: the live selfie is matched against an enrolled face, GPS confirms the person is inside the work zone, and the check-in is recorded with a tamper-resistant audit line. Biometric where it matters, without the hardware — and with a human confirming any decision the AI suggests.

Face as the biometric

A live selfie at check-in is matched against the enrolled photo, server-side. No physical reader to share, break, or replace.

GPS as the location check

Geofencing confirms the person is inside the site, post, ward, or branch at the moment of check-in. Outside the zone is flagged.

Each person uses their own phone

No queue at a shared device, no public touch surface. Works on any modern smartphone.

Anti-spoofing built in

Liveness checks reject printed-photo and screen-replay tricks. Impossible-GPS-speed and device-change signals are flagged, not silently accepted.

Anomaly detection

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

AI suggests, humans decide

Smart suggestions on regularisation and anomalies. Every consequential decision still needs a manager's click. No black-box automation.

Configurable retention

Biometric records auto-delete after the configured retention window. Audit history is kept separately, without the biometric image.

PDPL-compliant by design

Explicit consent on first use, encrypted storage, and data-subject rights (access, correction, erasure) through the staff app.

The biometric flow, end to end

01

Enrol once

The staff member is photographed in good light during onboarding. The photo is encrypted; consent is captured and timestamped.

02

Check in from any zone

They open the app at the work location, take a live selfie, and confirm. GPS + face match run together in seconds.

03

Live audit recorded

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

04

Anomalies surfaced

Out-of-zone, low-confidence, or unusual patterns are flagged for a human to decide — never silently accepted.

05

Auto-deletion at expiry

Biometric records auto-delete after the configured retention period; an audit summary is retained without the image.

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, retention configurable. Data-subject rights supported through the staff app.

No shared touch surfaces

Each person uses their own phone — no hygiene risk, no queue at a shared device, no recurring sanitisation overhead.

No third-party AI provider sees biometric data

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

Tamper-resistant evidence

Each check-in records match confidence, GPS, device, and time — audit-ready evidence for UAE Labour Law working-hour requirements.

Frequently asked questions

Is face matching really biometric for compliance?
Yes — facial geometry is a recognised biometric category. UAE PDPL treats it as biometric data and requires explicit consent, encryption, and retention controls. Aiya applies all three by default.
How accurate is the face 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.
What stops printed-photo or video spoofing?
Liveness checks reject obvious replay attempts. Combined with the GPS zone check and device-binding signals, the combination is harder to spoof than a fingerprint scanner that accepts any finger pressed against it.
How long is biometric data kept?
The retention period is configurable per organisation and aligned with your PDPL policy. The audit trail (who, when, where, confidence) is kept separately without the biometric image, preserving the record without holding personal data indefinitely.
Does the AI make decisions on its own?
No. The AI verifies identity and flags anomalies, but every consequential decision — approving a correction, accepting a flagged check-in — requires a human. No black-box automation.

See a live check-in — no kiosk

In a 20-minute demo we will enrol a face on a phone, run a live check-in, show the audit trail, and walk through the PDPL consent flow. No hardware involved.