Identity Verification in Digital Onboarding

Identity Verification in Digital Onboarding: Methods, Fraud Types & Integration

Table of Contents
Share this article

Identity verification is the part of digital onboarding that proves an applicant is a real, live person who owns the document they present and is not on a sanctions list. This guide explains the methods, the fraud each one stops, how to wire them into a sign-up flow with Verifik’s API, and a checklist for a fraud-resistant launch. Updated November 2026.

Identity verification in digital onboarding definition

What identity verification means in digital customer onboarding

In a branch, an employee looks at an ID and at the person holding it. Online, software has to do the same work with three questions: Is the document genuine and are its data correct? Is the person in front of the camera alive and the same person on the document? Is that identity allowed to become a customer, given sanctions, watchlists and your own fraud history? Every method below answers one of those questions.

You can try what this identity verification in digital onboarding guide covers in your own account: create a Verifik account.

Why onboarding is the main target for identity fraud in 2026

Onboarding is where a fraudster gets a new account, credit line or SIM card, so it attracts the most sophisticated attacks. Entrust’s 2026 Identity Fraud Report links deepfakes to about one in five biometric fraud attempts, says deepfaked selfies rose 58% in 2025 and puts injection attacks up about 40% year over year (Entrust). In Latin America, Unico and Liminal’s 2026 analysis of attacks between June 2025 and April 2026 found that sophisticated attacks combining deepfakes, injection and physical artefacts made up 23.3% of attempts overall and close to half in the region (Unico).

Most common types of identity fraud in digital onboarding

Fraud typeHow it worksCheck that stops it
Stolen identityReal person’s data and ID photo used by someone elseLiveness + 1:1 face match to the ID portrait
Synthetic identityReal ID number combined with a fake name, photo or addressRegistry validation of number and name together
Forged or altered documentsEdited photo, number or dates on a real templateDocument checks + registry validation
Presentation attacksPrinted photo, screen replay or mask in front of the cameraPassive liveness (PAD)
Deepfakes and injectionSynthetic video fed into the app, bypassing the cameraLiveness plus injection-attack detection and device integrity
Multi-accountingSame person opens many accounts with different documents1:N face search across your enrolled users
Money mules and sanctioned personsGenuine identity used to move illicit fundsSanctions, PEP and watchlist screening
Account takeover after onboardingAttacker changes the phone or email on an existing accountFace re-authentication against the enrolled template

Key identity verification methods used in digital onboarding

Identity verification methods used in onboarding

Document verification and OCR

The user photographs their ID; OCR extracts the fields and checks the format, dates and check digits. Scan Studio v3 (POST /v3/ocr/scan-studio) reads documents such as Colombia’s CC and RUT, passports (PA), DNI and Venezuela’s CCVE; v2 is retired on 25 November 2026 (docs). See OCR document verification.

Biometric verification: liveness and face match

Liveness confirms a live person is present; Verifik’s PAD is iBeta Level 2 certified and aligned with ISO 30107, with a default pass score of 0.6 (docs). The 1:1 face match compares the selfie with the ID portrait using a threshold from 0.67 to 0.95, default 0.85 (docs). Background in biometric verification.

Database and registry validation

The document number and name are checked against the issuing authority, which catches synthetic identities that a perfect-looking document would not reveal. Verifik validates identity documents in 20 countries (guide). For the business case, see database screening in onboarding.

OTP and contact verification

A one-time code by SMS or WhatsApp proves the applicant controls the phone they gave you, which is also the channel you will use for account recovery. See OTP verification for onboarding.

Sanctions, PEP and watchlist screening

Screening runs on every approval and again periodically. The OFAC check accepts a document number or a full name and returns whether the person is on the SDN list.

How to integrate identity verification into your onboarding flow

A common pattern is to create the onboarding session first, then run the checks server-side in a fixed order so a failed cheap check avoids an expensive one. Start with an app registration for the applicant:

curl -X POST "https://api.verifik.co/v2/app-registrations" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $VERIFIK_TOKEN" \
  -d '{
    "project": "<PROJECT_ID>",
    "projectFlow": "<PROJECT_FLOW_ID>",
    "email": "[email protected]",
    "phone": "3001234567",
    "countryCode": "+57",
    "fullName": "Ana Gómez"
  }'

Then run document, liveness, face match and screening from your backend. This Python orchestrator returns a decision and the evidence IDs you need for audit:

import os, requests

API = "https://api.verifik.co"
H = {"Content-Type": "application/json",
     "Authorization": f"Bearer {os.getenv('VERIFIK_TOKEN')}"}

def post(path, body):
    r = requests.post(API + path, json=body, headers=H, timeout=60)
    return r.status_code, r.json()

def ekyc(id_front_b64, id_portrait_b64, selfie_b64, document_type="CC"):
    # 1. Read the ID (Scan Studio v3; v2 is retired on 25 Nov 2026)
    code, doc = post("/v3/ocr/scan-studio", {"image": id_front_b64, "documentType": document_type})
    if code != 200:
        return {"decision": "retry_capture", "step": "ocr", "detail": doc}
    number = doc["data"].get("documentNumber", "").replace(".", "")

    # 2. Liveness on the selfie (PAD, default threshold 0.6)
    code, live = post("/v2/face-recognition/liveness",
                      {"os": "ANDROID", "image": selfie_b64, "liveness_min_score": 0.6})
    if code != 200 or not live["data"]["passed"]:
        return {"decision": "review", "step": "liveness", "evidence": live.get("id")}

    # 3. 1:1 face match: selfie vs. ID portrait (threshold 0.67-0.95, default 0.85)
    code, match = post("/v2/face-recognition/compare",
                       {"probe": [selfie_b64], "gallery": [id_portrait_b64],
                        "search_mode": "ACCURATE", "compare_min_score": 0.85})
    if code != 200 or match["data"]["score"] < 0.85:
        return {"decision": "review", "step": "face_match", "evidence": match.get("id")}

    # 4. Sanctions screening (OFAC SDN)
    r = requests.get(API + "/v2/ofac", headers=H, timeout=60,
                     params={"documentType": document_type, "documentNumber": number})
    if r.ok and r.json()["data"].get("foundInOFAC"):
        return {"decision": "reject_and_escalate", "step": "sanctions", "evidence": r.json().get("id")}

    return {"decision": "approve", "evidence": [doc.get("id"), live.get("id"), match.get("id")]}

Screening on its own, for example in a periodic re-check, is a single call:

curl -G "https://api.verifik.co/v2/ofac" \
  -H "Accept: application/json" \
  -H "Authorization: Bearer $VERIFIK_TOKEN" \
  --data-urlencode "documentType=CC" \
  --data-urlencode "documentNumber=80251972"

For a single call that compares and then checks liveness, use POST /v2/face-recognition/compare-with-liveness (docs).

Benefits of integrating identity verification into onboarding

  • Fewer fraud losses: stolen and synthetic identities are stopped before credit or services are granted.
  • Faster approvals: automated checks replace manual document review for most applicants.
  • Compliance evidence: each check returns a record you can show to a regulator or auditor.
  • Lower review cost: humans look only at exceptions, with the evidence attached.
  • Reusable identity: the enrolled face template lets you re-authenticate the customer later without new documents.

Identity verification use cases by industry

IndustryMain riskPriority checks
Banks and fintechStolen and synthetic identities, mule accountsLiveness, face match, registry, sanctions
TelecommunicationsSIM fraud and impersonationFace match at activation; Colombia’s Ley 2573 adds anti-impersonation duties
E-commerce and marketplacesFake sellers, chargebacksDocument + selfie for sellers, 1:N for repeat offenders
Healthcare and insuranceBenefit and claim fraudRegistry validation and re-authentication at claim time
Mobility and gig economyAccount sharing among drivers or couriersSelfie re-checks against the enrolled face

Identity verification best practices for digital onboarding

For banks, the full digital onboarding flow (regulatory requirements, steps and risk-based depth of checks) is in our bank eKYC guide.

Checklist for fraud-free digital onboarding

  • Use passive liveness on every selfie; do not rely on face match alone.
  • Protect the capture: in-app camera, device integrity checks and injection-attack detection.
  • Validate the document number and name against the official registry, not only the image.
  • Run 1:N face search to stop the same face under different names.
  • Screen sanctions, PEP and watchlists at onboarding and on a schedule.
  • Layer checks by risk; step up instead of rejecting on a single weak signal.
  • Log the evidence ID of every check and keep it for the retention period your regulator requires.
  • Offer an alternative path for people who cannot complete biometrics, as Colombia’s SFC requires.
  • Review false rejects and confirmed fraud monthly and retune thresholds.

For the user-experience side of the same flow, see digital onboarding best practices; for the full process, what is digital onboarding.

Verify identities during onboarding with smartENROLL

smartENROLL combines OCR, liveness, face match, registry validation, OTP and screening in one flow, hosted or via API, and returns signed evidence for every applicant. Book a demo to test it with your own documents.

Miguel Treviño - Co-founder and CTO, Verifik

Frequently Asked Questions (FAQs)

➖

What is identity verification in digital onboarding?

It is the set of checks that prove an online applicant is a real, live person who owns the ID they present and is allowed to become a customer: document verification, liveness, face matching, registry validation and sanctions screening.

➕

What are the most common types of onboarding fraud?

Stolen identities, synthetic identities, forged documents, presentation attacks with photos or masks, deepfakes and injection attacks, multi-accounting, money mules and account takeover after onboarding.

➕

How do you prevent identity fraud in digital onboarding?

Combine passive liveness, a 1:1 face match to the ID, registry validation of the document number and name, 1:N face search for duplicates and sanctions screening, and route weak signals to step-up or review.

➕

How do I integrate identity verification into my onboarding flow?

Create an onboarding session, then call the document, liveness, face match and screening APIs from your backend in order, store each evidence ID and return approve, step-up or review to your app.

➕

Can face matching alone stop deepfakes?

No. Face matching only compares two faces. You also need liveness detection against presentation attacks and controls against injection, where synthetic video is fed into the app without a real camera.

Explore more articles