The standards manual for VeriCam verification.
Twelve operational standards — ST-01 through ST-12 — covering intake, primary-source contact, officer review, supervisor approval, report issuance, revocation and re-verification.
Verification Methodology
- ST-01Verification principles
- ST-02Primary-source
- ST-03Officer review
- ST-04AI controls
- ST-05Outreach
- ST-06Evidence & audit
- ST-07Outcome framework
- ST-08Report integrity
- ST-09Fraud & risk
- ST-10Security architecture
- ST-11Privacy & consent
- ST-12Limitations & scope
Verification principles.
Eight principles that govern every decision the platform supports. They are enforced through both organisational practice and platform code.
Primary-source confirmation
Every verified report is anchored to a response from the issuing Cameroonian institution — either through their onboarded portal or documented manual outreach.
Reports cite the channel of record (portal / email / phone / letter / in person), the person contacted, and the response date.
- No verified outcome without institution response
- Channel + person + date stored on every InstitutionResponse
- Suspected-fraud signals halt issuance
Human-reviewed decisions
Final outcomes are determined by trained verification officers, not by automated systems. Officers compare evidence against institutional response.
An officer cannot self-issue. A supervisor must approve before any report enters circulation.
- Officer classification with written rationale
- Supervisor approval enforced in code
- Two-person sign-off recorded in the audit log
Evidence-based classification
Outcomes are classified strictly on the evidence on the case file: documents, AI extraction, institution response, officer notes.
Officers cite specific evidence when classifying. The rationale is stored verbatim on the report payload.
- Rationale required for every classification
- Evidence references stored by ID
- Reports carry the rationale text
Controlled AI assistance
AI extracts structured data and surfaces anomalies. It accelerates analysts — it never decides outcomes.
Every AI extraction is shown with per-field confidence and detected anomalies. Officers can override any AI inference.
- Provider + version recorded per extraction
- Confidence + anomalies always visible
- AI signals are advisory only
Audit preservation
Every action that touches a case is recorded with actor, timestamp, IP, user agent and a diff of the change.
AuditLog rows are never overwritten. Reports can be reconstructed against the audit trail years after issuance.
- Immutable AuditLog table
- Append-only entries
- Diff stored for state transitions
Institutional transparency
The methodology, the institutions tracked, and the verification outcomes are explicit. We do not approximate; we do not infer.
Where an institution cannot be reached, the outcome is Unable to Verify — never an implied positive.
- Published standards
- Disclosed contact channels
- Outcome definitions binding and public
Re-verifiable reports
Every report carries a verification ID, machine-checkable metadata and a public re-verification page so recipients can re-confirm status indefinitely.
Payload snapshots are immutable. Revocation or supersession is reflected immediately on the public page.
- Snapshot signed with payload hash
- Public verification ID never recycled
- Revocation status surfaced publicly
Explicit scope & limitations
Reports state what was verified, by whom, with what evidence, and what limitations apply.
VeriCam does not award qualifications, accredit institutions, or evaluate equivalency. Reports disclose this scope.
- Limitations footer on every report
- Disclaimer included in the audit payload
- Equivalency requests routed externally
Primary-source verification methodology.
Verified outcomes are grounded in a response from the issuing institution. We carry the evidence — the verification itself belongs to the institution.
- ·Routing. Onboarded institutions receive cases in their portal. Others receive documented outreach.
- ·Channel of record. Portal / email / phone / letter / in person — captured and disclosed on the report.
- ·Evidence capture. Response, contact person, date and uploaded supporting material attached to the case.
- ·No-response handling. After two documented attempts across different channels, the outcome is Unable to Verify.
Match against the institutional intelligence database (aliases, acronyms, French/English names).
Onboarded → portal. Not onboarded → manual outreach with channel-of-record discipline.
Channel, person, email/phone used, attempt count, response date — all stored on the case.
Institution response stored verbatim. Officer ties it to the case as InstitutionResponse.
Officer compares submitted credential vs response; classifies one of five outcomes; writes rationale.
Officer review standards.
A trained verification officer interprets the evidence and classifies the outcome. Officers cannot self-issue: a supervisor approval is enforced before any report enters circulation.
- ·Officer qualifications. Trained on the standards manual; periodic recalibration on new credential templates.
- ·Rationale requirement. Every classification requires a written rationale citing specific evidence.
- ·Escalation thresholds. High-risk and suspected-fraud cases automatically route to a senior officer.
- ·Supervisor approval. Two-person sign-off enforced in code; no self-approval is possible.
| Field | Submitted | Institution | Match |
|---|---|---|---|
| Applicant | Marie Tchakounté | Marie Tchakounté | Match |
| DOB | 04 Mar 1996 | 04 Mar 1996 | Match |
| Programme | BSc Computer Science | BSc Computer Science | Match |
| Year | 2022 | 2022 | Match |
| Grade | Second Class Upper | Second Class Upper | Match |
| Certificate № | UB-BSC-2022-3811 | UB-BSC-2022-3811 | Match |
AI-assisted intake controls.
AI extracts structured data, flags anomalies and surfaces risk signals. It accelerates officers — it never issues verifications, classifies outcomes or contacts institutions on its own.
- Extracting structured fields from documents.
- Surfacing per-field extraction confidence.
- Detecting risk signals (template mismatch, altered image, etc.).
- Matching the issuing institution against the alias database.
- Flagging duplicate certificate numbers.
- Prioritising cases by SLA and risk for the officer queue.
- Issuing a final verification outcome.
- Classifying a case as Verified, Not Verified or anything else.
- Contacting institutions on behalf of VeriCam.
- Approving or revoking reports.
- Communicating with applicants or requesters.
- Overriding a documented institutional response.
Institutional outreach standards.
When an institution is not yet onboarded, our officers conduct documented outreach. The channel, contact, attempt count and response are captured per attempt.
| Channel | Target | Standard SLA | Evidence retained | Attempts | Confidence |
|---|---|---|---|---|---|
| Portal | Onboarded institutions | 2 business days | Confirmation row in InstitutionResponse, optional uploaded letter | 1 | High |
| Registrar address of record | 5 business days | Header, body, attachments — full thread retained | ≤ 2 | High | |
| Phone | Registrar / exam office | Same day | Officer's contact log + named person | ≤ 2 | Medium |
| Letter | Ministry-recognised postal address | 10 business days | Sent receipt + scanned response | 1 | Medium |
| In person | Registrar office | Per appointment | Officer's site visit log + counter-signed evidence | 1 | High |
Where digital channels do not respond within the documented attempt budget, our Cameroon-based team performs additional follow-up — directly with the institution's registrar's office. Every follow-up is documented on the case file with actor, channel, timestamp and outcome.
- Telephone follow-up. Voice contact via the institution's published numbers. Call attempts and notes are recorded.
- Written institutional letter. Formal letter to the registrar's office through the institution's official mailing channel.
- Registrar office visit. In-person visit to the institution's registrar / academic registry, by a named officer.
- Documented evidence. Visit log, signed receipt, photographed records or registrar attestation — attached to the case file.
Evidence & audit trail.
Every action that touches a case writes an immutable audit row. Reports can be reconstructed against the audit trail years after issuance — including the exact evidence that supported the decision.
- ·Append-only. AuditLog rows are never overwritten — superseded by new entries when state changes.
- ·Diff-aware. State transitions store before / after values; document hashes link evidence.
- ·Actor-attributed. Every row carries actor identity, role, IP and user agent.
- ·Long-horizon retention. Cases retained per published retention policy; legal hold supported.
- 2026-05-03 09:14:02ISSUESolange MbargaSuper AdminReport RPT-CMR-2026-000142-A · sha256=8f3c…2b4a
- 2026-05-03 09:13:51APPROVESolange MbargaSuper AdminVerificationCase → status=VERIFIED
- 2026-05-03 08:58:32CLASSIFYEtienne NkomoOfficerOutcome=VERIFIED · rationale 217 chars
- 2026-05-03 08:42:09RESPONDUB RegistrarInstitutionInstitutionResponse · channel=PORTAL
- 2026-05-03 08:31:14OPENUB RegistrarInstitutionReviewed evidence bundle (1 doc, 2.4 MB)
- 2026-05-03 08:16:00UPLOADEtienne NkomoOfficerInstitutionEvidence · evidence-bundle.pdf
- 2026-05-03 08:14:21EXTRACTSystemAI PipelineAIExtraction · 8 fields · 1 risk signal
- 2026-05-03 08:13:02ASSIGNSolange MbargaSuper AdminOfficer = Etienne Nkomo
- 2026-05-03 08:12:44CREATEHarvard · Sarah WhitmanForeign requesterVerificationCase CMR-2026-000142
Verification outcome framework.
Every case resolves into exactly one of five outcomes. Each outcome has a precise institutional meaning, an evidence threshold, an explicit limitation and a recommended interpretation.
Primary-source confirmed all material facts.
Institution response confirming applicant identity, credential type, year, classification and certificate number.
The issuing Cameroonian institution explicitly affirmed the credential as presented.
Does not warrant equivalency or accreditation status. Reports remain subject to revocation if facts change.
Accept the credential subject to your own admissions, employment or licensing policy.
Some facts confirmed; others could not be confirmed.
Institution confirmed core enrolment / award but could not validate every material claim (e.g. classification).
The verified subset is supported by primary source; the unverified subset is not.
Officer rationale specifies which facts were and were not confirmed.
Decide whether the confirmed facts are sufficient for your decision. Request additional evidence if not.
After documented good-faith outreach, no response.
At least two documented outreach attempts across different channels with no authoritative response.
This is not a finding of fraud — only that the issuing institution did not respond.
The case file records every outreach attempt with channel, contact and date.
Request additional documentation directly from the applicant, or escalate to your evaluator.
Primary source disowns the credential as presented.
Institution response explicitly contradicting the document submitted, with specific reasons stated.
The credential cannot be relied upon as the applicant presented it.
Officer rationale specifies the nature of the institution's disavowal.
Treat the credential as unconfirmed. Consider downstream policy (rejection, follow-up, audit).
Material indicators of fabrication, evidence-backed.
Institution disavowal combined with concrete document indicators (forged seal, altered image, fabricated cert number).
Reserved for cases where evidence rises to the level of an integrity concern; reported to compliance.
Officer rationale references the specific anomalies and the institution statement.
Halt processing. Reference the report rationale and consider regulatory disclosure obligations.
Report integrity & re-verification.
Every report is a state machine. The verification ID is immutable; the report status reflects the current trust state. Recipients re-check status before relying on any report.
- ·Immutable payload snapshot. Report contents are frozen at issuance; cryptographic hash linked to the case.
- ·Public re-verification. QR + verification ID resolve to a public page disclosing status only — never documents.
- ·Revocation. When new evidence invalidates a report, status moves to REVOKED with a recorded reason.
- ·Supersession. When a corrected report is issued, the prior version is marked SUPERSEDED with a link to the successor.
Report generated, payload snapshot hashed, verification ID minted.
Default state. Public lookup returns VALID + outcome + dates.
Past the report's expiry date. Still verifiable historically; recipient should re-verify.
New evidence invalidated the report. Reason recorded; public page shows REVOKED prominently.
A corrected report has been issued; this version links forward to its successor.
Every report carries a public verification ID. The public page resolves the ID to the current state of the report — never the historical state. This ensures that revocation and supersession are reflected immediately for any recipient who re-checks.
Fraud detection & risk controls.
Eight defined fraud signals feed a risk score. High signals trigger escalation; suspected-fraud halts issuance and routes to compliance review.
Document layout, header, font or seal placement does not match the institution's known template family.
Indicators of digital alteration (resampling, inconsistent compression, edge artefacts, overpainting).
Certificate number previously seen on another case with a different applicant.
Issuance, graduation, signing or stamp dates internally inconsistent or chronologically impossible.
Issuing institution outside the recognised Cameroonian registry, with no accreditation reference.
Resolution or contrast below thresholds — increases extraction error rate; not a fraud indicator alone.
Institution response that materially contradicts the submitted document.
Applicant, institution or certificate number appears on a prior confirmed-fraud case.
Security & compliance architecture.
The same controls that make a verification credible — least privilege, encrypted storage, audit retention, consent — are enforced in code rather than promised in copy.
Uploads encrypted across the pipeline. Production deployments use a managed KMS.
Applicants explicitly authorise release to a named recipient; consent is logged and revocable.
Each report carries an immutable verification ID, payload snapshot and public QR.
Six roles with distinct permissions. Officers cannot self-issue reports.
Risk signals fire alerts. Suspected fraud halts issuance and routes to compliance review.
AuditLog rows are never overwritten; full history is preserved across the case's lifetime.
Public verification pages reveal status and dates only — never documents or PII.
Two-person sign-off is enforced in code, not policy. Supervisors cannot self-approve their own decisions.
Privacy & consent.
Applicants explicitly authorise the release of their credentials to a named recipient. Consent is logged on the case, can be revoked prior to issuance and is reflected on the report.
Limitations & disclosures.
A VeriCam report confirms a specific fact established against specific evidence. It is not a substitute for accreditation or equivalency evaluation. The following are out of scope.
We do not certify that the issuing institution is accredited by any specific body. Accreditation is a separate, time-bound determination.
We do not evaluate the Cameroonian credential against another country's framework. That is the domain of credential evaluators (e.g. WES, ENIC-NARIC).
A report reflects status at issuance. New evidence may move the report to REVOKED or SUPERSEDED.
We may verify a credential as Verified even if the submitted document itself is poor quality, provided the institution confirms the underlying record.
Credentials issued outside Cameroon are out of scope. Reach out to your regional credential evaluator.
Public re-verification reveals status only — never documents or PII. Recipients with case access can see full details via their dashboard.
Hold us to the standards on this page.
Every report we issue can be traced to the standards above. Submit your first case, read the operational workflow, or apply to onboard your institution.