Sigilith Product
Sigilith Sentry

AdvancedeKYCandfrauddecisioningforregulatedonboarding.

Sigilith Sentry combines document intelligence, live identity verification, AI-assisted adjudication, and human review for institutions that need fast, controlled, and defensible onboarding decisions.

3Verification Stages
3Decision Layers
100%Audit Trail

Workflow Overview

01
Capture
Document submission
02
Verify
Liveness & face match
03
Adjudicate
Fraud assessment
04
Escalate
Human review

How It Works

Three-stage verification. One defensible decision.

Document intelligence, live identity verification, and fraud adjudication in a single controlled workflow.

01
01 / STAGE

Document Intelligence

Capture and extract identity data from complex documents using advanced OCR, including support for difficult scripts and document formats such as Bangla NID workflows.

02
02 / STAGE

Live Identity Verification

Collect a reference selfie, match identity signals, and confirm live presence through active liveness prompts such as directional head movement.

03
03 / STAGE

Fraud Adjudication and Escalation

Suspicious or failed cases move through automated checks, AI-assisted re-review, and human escalation when final judgment requires operator oversight.

Capabilities

Built for fraud-heavy onboarding environments.

Six capabilities that go beyond standard identity verification.

DOCUMENT_OCR
01

Advanced OCR for complex identity workflows

Extract structured data from documents that standard OCR fails on, including non-Latin scripts, worn documents, and Bangla NID formats.

FACE_MATCHING
02

Reference selfie and face matching

Capture a reference image and run identity matching against document data to confirm the applicant is who they claim to be.

LIVENESS
03

Active liveness prompts

Directional liveness challenges confirm the applicant is physically present at the time of verification, not submitting a static image.

FRAUD_DETECTION
04

Fraud detection and rejection handling

Automated signals detect document manipulation, spoofing attempts, and inconsistencies before cases reach a human reviewer.

SENTINEL_REVIEW
05

AI-assisted second-pass review

Sentinel Review performs a second-pass assessment on rejected or ambiguous cases, improving accuracy and throughput before human escalation.

DECISION_CONSOLE
06

Human-in-the-loop final decisioning

Operators resolve edge cases in Decision Console with full access to extracted data, fraud signals, liveness results, and case history.

Sentinel Review

AI-assisted review before human escalation.

Sentinel Review performs a second-pass assessment on rejected or ambiguous onboarding attempts, helping teams identify cases that deserve deeper review before a final human decision is made.

01
AUTO CHECK
Initial screening
02
SENTINEL
AI second pass
03
OPERATOR
Human decision
ACCURACY

Reduces false rejection rate

Second-pass assessment catches valid cases that automated checks flag incorrectly, before they reach a human operator.

THROUGHPUT

Improves review efficiency

AI pre-triage means operators spend time on genuinely ambiguous cases, not routine reviews that could be resolved automatically.

CONTROL

Human judgment retained

Sentinel Review informs, it does not decide. Final approval or rejection remains with the operator.

Decision Console

Operational review for high-risk cases.

Rejected or suspicious cases are routed into the Decision Console, where teams can inspect extracted identity data, reference imagery, liveness outcomes, fraud signals, and AI review results before resolving the case.

Review extracted data

Inspect structured identity fields pulled from the document alongside the original captured image.

Inspect fraud signals

View automated fraud scores, liveness results, and any flags raised during AI adjudication.

Track escalation history

See how a case moved through automated, AI-assisted, and human review stages.

Resolve edge cases with operator control

Approve, reject, or request additional information with a full audit trail of the decision recorded.

DECISION_CONSOLE
ESCALATED
Case Summary
DOCUMENT_TYPENational ID
LIVENESS_RESULTPass
FRAUD_SCORE0.00Clear
SENTINEL_VERDICTAmbiguous
ESCALATED_BYAutomated pipeline
Operator Actions
Approve
Reject
Request Info

Governance

Designed for controlled decision workflows.

Sigilith Sentry is built for environments where decisions must be defensible, traceable, and subject to human oversight.

Human oversight

Unresolved or high-risk cases require operator resolution before a final decision is issued.

Traceable case progression

Every case carries a record of automated checks, AI verdicts, and operator actions from start to resolution.

Structured escalation workflow

Cases move through defined stages. Nothing is silently dropped or auto-approved without a traceable reason.

AI-assisted, human-decided

AI-assisted review is designed to support, not replace, final human judgment in regulated onboarding decisions.

Enterprise-ready deployment

Deployment posture is designed for institutional environments with existing compliance, security, and access control requirements.

See it in action

See Sigilith Sentry in action.

Request a live walkthrough of our advanced eKYC and fraud decisioning workflow.