Centeem applies a strict framework to combat money laundering and terrorist financing, in accordance with the obligations imposed by the Bank of Algeria (Instruction 06-2025) and the Algerian anti-money laundering code.
1. Fundamental principles
As a Payment Service Provider (PSP) undergoing licensing, Centeem is subject to the same AML/CFT obligations as banks:
- Due diligence towards each user (identification, verification, ongoing monitoring)
- Mandatory reporting to the CTRF (Financial Intelligence Processing Unit) for transactions reaching the legal thresholds
- Retention of KYC data and transaction history for 5 to 10 years
- Continuous training of the internal team on new money laundering schemes
2. KYC — Identity verification (Article 102)
No Centeem account is opened without a complete KYC process:
- Official identity document: Algerian biometric national ID card (with secure chip) or passport for foreign visitors
- Selfie with liveness detection: anti-photo, anti-video, anti-mask algorithm
- Valid visa for foreign users
- International sanctions screening: automatic checking against OFAC, EU, UN lists
- Duplicate detection: the same individual may open only one Personal account and one Pro account
KYC is renewed annually or biennially depending on the account tier (limit level).
3. Automatic reporting thresholds
In accordance with Articles 88-90 of Instruction 06-2025:
| Type | Threshold | Automatic action |
|---|---|---|
| Single CTR (Currency Transaction Report) | ≥ 1,000,000 DZD | Immediate report to the CTRF |
| Monthly cumulative CTR | ≥ 10,000,000 DZD/month on an account | Immediate report to the CTRF |
| STR (Suspicious Transaction Report) | Suspicious pattern detection | Monthly report to the CTRF |
4. Suspicious patterns detected (Article 91)
Our internal fraud + AML engine automatically detects the following patterns:
- Structuring: deliberate splitting of large amounts into several small transactions to avoid CTR thresholds (e.g. 10 × 99,000 DZD instead of 1 × 990,000 DZD)
- Smurfing: use of multiple accounts to disperse flows
- Fan-in pattern: abnormal convergence of funds toward a single account
- Excessive velocity: number of transactions/hour far exceeding the user profile
- Shadow merchant: a personal account used to collect invisible commercial activity
- Abnormal nighttime activity: large transactions between 00:00-05:00
5. Precautionary measures
In the event of suspicious activity detection, Centeem may, without notice:
- Temporarily block a transaction pending compliance review
- Freeze the account until clarification (the user is notified with a generic reason)
- Request supporting documents on the origin of the funds
- Reject a top-up if the source is deemed non-compliant
- Close the account as a last resort, returning the balance after verification
6. Your rights
If your account is subject to a precautionary measure:
- You receive a notification with a reason (usually generic to preserve the investigation)
- You may provide supporting documents via the contact form
- You may refer the matter to the complaints service or to the BA Payment Services Mediator
- Legal recourse remains possible after exhausting amicable channels
7. Confidentiality of reporting
If Centeem issues a CTR or STR concerning you, we are not allowed to inform you of it. This is the legal principle of « tipping-off » which prohibits alerting a person under investigation. The CTRF then decides on the follow-up (closure, request for additional information, judicial transmission).
8. External oversight
Our AML framework is audited every year by:
- An independent external audit firm (report submitted to the BA)
- The Bank of Algeria (unannounced inspections possible)
- Our internal Compliance Committee (with an independent member)
9. Compliance contact
For any question about our AML framework, write to compliance@centeem.com. Our compliance officer responds within 15 business days (legal deadline for formal requests).
This document is a summary of our AML policy. The full framework (internal procedures, team training, scoring models) is confidential for security reasons — revealing our methods would weaken their effectiveness against fraudsters.