Workflow teardown

AML/KYC Refresh Workflow

A workflow for refreshing AML/KYC evidence, ownership details, screening status and approval history.

Workflow teardown

Trigger

A client, staff member or compliance event starts the workflow.

Collect

HubSecure collects required information, files and approvals securely.

Review

Owners review status, evidence, risk and next steps from the client record.

Prove

The final record shows requests, uploads, messages, decisions and approvals.

QuestionWhat to inspect
Who owns the next step?Task owner, due date and blocker status.
What is missing?Required documents, client responses and approvals.
What proof exists?Upload source, review notes, permission history and approval trail.

Evidence created at each refresh step

Every action in the AML/KYC refresh creates a timestamped evidence entry. The complete trail satisfies regulatory record-keeping requirements without manual logging.

Refresh triggered
Trigger source, client ID and due date logged
Documents requested
Request type, channel, deadline and request ID logged
Identity verified
Document OCR result, liveness check and face match score logged
Screening run
Screening source, result and timestamp logged
Risk assessed by AI
AI risk score, model version and confidence logged — human review required
Human approval granted
Approver identity, decision and timestamp logged

Sentinel Identity: document OCR, liveness and face match

The AML/KYC refresh workflow integrates with Sentinel Identity to automate the identity verification step with governed AI — reducing manual document checking without removing human oversight.

Document OCR

Sentinel Identity reads identity documents automatically — extracting name, date of birth, document number and expiry — and flags discrepancies against the client record.

Liveness detection

A liveness check confirms the document holder is a real, present person — not a photo or screen replay. The result is logged with a confidence score and timestamp.

Face match

Sentinel compares the document photo to the liveness capture and returns a match score. Low-confidence matches are automatically escalated for human review.

Governed AI for risk assessment

AI assists with risk assessment during the refresh — but never makes the final call. Every AI-generated risk score requires human review and explicit approval before the client record is updated.

How governed AI works in this workflow

The AI Operator reviews the refreshed client data, screening results and identity verification output to produce a risk score and a plain-language summary of the key factors. The score is presented to the compliance officer alongside the evidence. The compliance officer reviews, can adjust the score with a documented reason, and then grants or withholds approval. No client status is changed until the human approves. Every AI action — the inputs used, the score generated, the model version and the timestamp — is logged to the Evidence Timeline.

Evidence Timeline integration

The complete AML/KYC refresh trail is written to the Evidence Timeline automatically. When an auditor or regulator requests proof of your refresh process, the timeline is searchable and exportable in one click.

Refresh triggeredAnnual refresh due — client notified via secure mail — 2026-05-21 08:00
Identity documents receivedPassport and proof of address uploaded via client portal — file hashes stored
Sentinel Identity verification completeOCR: pass — Liveness: pass — Face match: 0.97 — no discrepancies flagged
AI risk assessment generatedRisk score: Low (0.18) — model: sentinel-risk-v4 — awaiting human review
Compliance officer approvedApprover m.chen reviewed AI score and evidence — approved — client record updated

Learn more about the Evidence Timeline →

Canonical hubs

Source-of-truth pages for this topic

These hub pages tell buyers and search engines how this page fits into the wider HubSecure information architecture.

Recommended next step

Continue the evaluation path

The next page should move the buyer from information to comparison, workflow review, template use or private rollout readiness.

Official references

Sources to verify the compliance context

HubSecure content is written for workflow evaluation, not legal advice. Use these official sources to verify regulatory and assurance context.