Computer Vision (iBeta Level 3) • CV-04
KYC Onboarding
Open the account remotely without opening a fraud channel.
Scope
What the engagement covers.
Remote onboarding must satisfy regulators and reject determined fraud while keeping abandonment low. This capability combines document authentication, liveness and face matching into one decisioned flow.
Included capabilities
- Identity document capture, OCR and authenticity verification
- Face-to-document matching with liveness confirmation
- Watchlist, sanctions and duplicate applicant screening
- Risk-based routing to straight-through approval or manual review
- Full decision evidence retained for regulatory inspection
Outputs and deliverables
- Onboarding flow design and decision rules
- Document and biometric integration configuration
- Manual review workflow and reviewer guidance
- Funnel, fraud and regulatory evidence reporting
Workflow
How it is delivered, step by step.
Each step has an owner, an entry condition and an artefact that has to exist before the next step begins.
01DesignFlow, assurance requirements, abandonment targets and review capacity.
02IntegrateDocument, liveness, matching and screening components.
03DecisionRisk-based routing rules with thresholds and referral criteria.
04PilotLive pilot measuring approval, abandonment and fraud detection.
05OptimizeThreshold and flow tuning against the measured funnel.
Use cases
Where this is typically applied.
Telecom SIM registration under regulatory identity rules
Marketplaces onboarding sellers at scale
Delivery model
The operating pattern for Computer Vision (iBeta Level 3).
The same delivery discipline applies across every capability in this line, so combined engagements stay coherent.
Define
Use case, lawful basis, accuracy targets and acceptance thresholds.
Prove
Bench test on representative data, including attack and edge-case sets.
Integrate
SDK or API embedding, gallery design and decision policy.
Deploy
Edge or on-premise rollout, tuning and human adjudication workflow.
Assure
Accuracy monitoring, bias review, revalidation and retention enforcement.
Integration
- Core banking, onboarding and case management platforms via SDK or API.
- Physical access controllers, turnstiles and video management systems.
- Identity registries and watchlists under a controlled matching policy.
- SIEM and audit platforms for decision evidence and administrative logs.
Engagement approach
Every deployment is scoped with a documented lawful basis, retention schedule and human adjudication path before a single template is enrolled. Accuracy and bias are measured on the client’s own population, not on vendor benchmarks.
Related
Other capabilities in Computer Vision (iBeta Level 3).
CV-01
Facial Recognition
1:1 verification and 1:N identification at gallery scale.
CV-02Liveness & Anti-Spoofing
Prove there is a live person in front of the camera.
CV-03Digital Identity
A reusable, governed identity credential instead of repeated verification.
CV-05Face, Body & Object Recognition
Understand what is in the scene beyond identity.
CV-06Vehicle Analytics
Read, classify and track vehicles automatically.
CV-07Smart City
Operational intelligence for public space at city scale.
CV-08Video Forensics
Find the thirty seconds that matter in ten thousand hours.
CV-09Border & Public Safety
High-assurance identity at controlled crossing points.
CV-10SDK / API Integration
Put biometric capability inside the applications people already use.