Biometric
Biometric Verification. Face + voice, B+V ≥ 80%.
Every biometric-flagged event (check-in, payout, sensitive access) triggers a face-match + voice challenge. Templates are encrypted at rest, never used in denominators for KPIs, and revocable on request.
Biometric
Unsplash · @aleksandarcvetanovic
Headline figure
92%
Compliance
vs 80% required
Every event on this module leaves an audit-trail row exportable as JSON.
92%
Compliance
vs 80% required
98.4%
Face-match confidence (median)
Live instance · rolling 30 days
AES-256
Encryption at rest
Every template
Automation
What Tshomi does automatically.
Face capture during check-in
Voice challenge for elevated actions
Template comparison + confidence scoring
Revocation endpoint for participants
Flow
Step by step.
1
Enrol
Face + voice sample captured once at onboarding.
2
Challenge
At every biometric-flagged event.
3
Match
Confidence score computed + logged.
4
Log
Audit-trail row: who, when, confidence.
1
Enrol
Face + voice sample captured once at onboarding.
2
Challenge
At every biometric-flagged event.
3
Match
Confidence score computed + logged.
4
Log
Audit-trail row: who, when, confidence.
Sample conversation
What the participant actually sees.
Tshomi · WhatsApp · in-language
Ndicela uthathe iselfie encinci — 3 seconds, no filters.