Regulation · Nigeria

Nigeria: CBN Baseline Standards for Automated AML Solutions

Nigeria’s central bank now sets the minimum for the systems that screen customers, monitor transactions and report suspicious activity. Every bank and financial institution the CBN supervises has to meet it: deposit money banks by 10 September 2027, everyone else by 10 March 2028. The standards run to 119 requirements. This page takes you through each one: the CBN’s own words, what an examiner will look for, and the detection rules that help you meet it.

Issued by
Central Bank of Nigeria
Reference
BSD/DIR/PUB/LAB/019/002
Issued
Applies to
All banks and other financial institutions supervised by the CBN
Implementation roadmap due
Full compliance, deposit money banks
Full compliance, other financial institutions
Updated

Detection rules for this regulation

25 rules in the library help meet 18 of its 119 requirements. Most of the others are about systems, governance and process, where no detection rule applies.

Every requirement

All 119, in the regulation’s own order. Open a part to read it.

About this page

This page lists every requirement in sections 5 and 6 of the standards and the fraud-monitoring duty in section 4(j), quoted from the CBN’s text, with what an examiner will look for, what meeting it can look like and, where one exists, a detection rule from this library that helps. The examiner notes draw on the CBN’s Guidance Note of 31 March 2026 and its Implementation Roadmap Template where they speak to the area.

Compliance is assessed for each institution, not for a system. The Guidance Note says that claims by vendors of “full compliance” or “alignment with CBN standards” are not recognised by the CBN. Where this page points to an Archer feature, it is an example of what a requirement can look like, not a claim that it is met.

4(j) Fraud detection and prevention

1 with rules 5 requirements

Section 4 (Scope) states this duty for every institution. The requirements of section 5.6 add to it where the AML solution is also used for fraud monitoring.

4(j) Automated fraud monitoring in every institution

Section 4(j) says:

Fraud detection and prevention: although this Standard does not mandate a unified AML–fraud system, financial institutions are required to deploy automated fraud-monitoring capabilities appropriate to their fraud-risk profiles. These capabilities may be hosted on the same platform as the AML Solution or on a separate fraud-management system, provided that:

What an examiner will look for
The Implementation Roadmap Template has a “Fraud Monitoring and Detection” row in its core table, with the current state, target state, actions, owner, dates and whether evidence is available (Template, section 3). Expect the automated fraud controls in use, which risks in the institution’s fraud-risk assessment each one addresses, and whether they run on the AML platform or a separate system.
What this can look like
Fraud rules run automatically on the channels the institution offers, chosen from its own fraud-risk assessment, whether in the AML system or a separate fraud system.

4(j)(i) Fraud and AML systems exchange risk signals

These capabilities may be hosted on the same platform as the AML Solution or on a separate fraud-management system, provided that:

both systems can exchange relevant risk signals seamlessly;

What an examiner will look for
Evidence that risk signals pass between the fraud and AML systems automatically, for example a confirmed fraud case reaching the customer’s AML risk profile, and AML alerts visible to fraud investigators. Section 5.6(b)(ii) asks for the first of these where one solution does both.
What this can look like
A confirmed fraud case raises the customer’s AML risk rating without anyone retyping it, and fraud investigators see a customer’s open AML alerts.

4(j)(ii) No blind spots, duplication or delays

These capabilities may be hosted on the same platform as the AML Solution or on a separate fraud-management system, provided that:

the architecture avoids blind spots, duplication, and delays in detection; and

What an examiner will look for
The roadmap asks for the current and target architecture, with diagrams, and how the target will “eliminate fragmentation” (Template, sections 4.1, 4.2 and 4.6). Expect a map showing that every channel’s events reach fraud or AML monitoring, which system handles what, and how quickly events pass between them.
What this can look like
A diagram where every channel feeds monitoring, no type of event is checked by neither system, and no alert is worked twice by two teams.

4(j)(iii) Fraud controls in real time or near real time

These capabilities may be hosted on the same platform as the AML Solution or on a separate fraud-management system, provided that:

fraud-specific controls meet real-time or near-real-time expectations, which AML transaction monitoring does not require.

What an examiner will look for
The glossary defines real-time monitoring as able to “trigger a pre-emptive action within stipulated timeframe that can prevent the completion of a suspicious activity (e.g., before funds are irreversibly settled)”. For each data source the roadmap must say whether it is real-time, near real-time or batch (Template, section 5.3). Expect measured response times for the fraud controls on each channel.
What this can look like
Fraud checks answer while the payment or login is still in progress, so the riskiest can be stopped before the money moves.

4(j), last paragraph Logical separation where one solution does both

Section 4(j) adds:

Where a single solution is used for both AML and fraud, the institution must ensure logical separation of rules, maintain the integrity of AML/CFT/CPF controls, and demonstrate that fraud-risk requirements are adequately addressed.

What an examiner will look for
Where one solution does both: the rule list showing which rules are fraud and which are AML, AML alert metrics tracked on their own, and evidence that the fraud risks in the fraud-risk assessment are covered. Section 5.1 asks for the same segregation of fraud capabilities on a shared platform.
What this can look like
Fraud and AML rules are labelled, owned and tuned separately in the one system, and AML detection is measured on its own so a fraud change cannot quietly weaken it.

5.1 AML Solution

3 with rules 13 requirements

5.1(a) Customer identification and verification

The AML solution shall, at a minimum, support:

Customer Identification and Verification (including remote/onboarding channels)

What an examiner will look for
The Guidance Note lists customer identification and verification among the functions an AML solution must support at a minimum, as part of “an integrated control environment” (Guidance Note, section 5). Expect to show how identity results from onboarding, including remote channels, reach the AML solution; the detail is tested under 5.2.
What this can look like
The results of identity checks at onboarding, in branch or remote, are stored on the same customer record that screening and monitoring use.

5.1(b) Customer risk assessment and profiling

The AML solution shall, at a minimum, support:

Customer Risk Assessment and Profiling

What an examiner will look for
Listed as a minimum function in the Guidance Note (section 5); the roadmap must show data flowing from “customer onboarding → risk profiling” (Template, section 4.4). Detail under 5.4.
What this can look like
Every customer has a risk rating the system computes and keeps current, visible wherever the customer is reviewed.

5.1(c) Sanctions and watchlist screening

The AML solution shall, at a minimum, support:

Sanction and Watchlist Screening

What an examiner will look for
Listed as a minimum function in the Guidance Note (section 5); sanctions and watchlist data are a minimum data source the roadmap must cover (Template, section 5.2). Detail under 5.3.
What this can look like
Customers and payment parties are screened against sanctions lists and the institution’s own watchlists, and matches go to an analyst.

5.1(d) PEP and high-risk customer screening

The AML solution shall, at a minimum, support:

Politically Exposed Persons (PEPs) and High-Risk Customer Screening

What an examiner will look for
The Guidance Note lists PEP screening among the minimum functions (“sanctions, watchlist, and PEP screening”, section 5); the high-risk customer part comes from 5.1(d) itself. PEP databases are a minimum data source in the roadmap (Template, section 5.2). Detail under 5.3(a)(vi).
What this can look like
Customers and beneficial owners are checked against PEP data, and PEPs and other high-risk customers are flagged for enhanced due diligence.

5.1(e) Transaction monitoring for ML/TF/PF

The AML solution shall, at a minimum, support:

Transaction Monitoring for ML/TF/PF risk

What an examiner will look for
Listed as a minimum function in the Guidance Note (section 5). The CBN assesses “the demonstrated effectiveness of the solution in supporting detection” (Guidance Note, section 7), so expect to show alert volumes and outcomes, not only that monitoring runs. Detail under 5.5.
What this can look like
Every relevant transaction passes through monitoring scenarios for money laundering, terrorism financing and proliferation financing typologies. An institution that offers virtual asset services also runs scenarios on its virtual asset transfers, such as chain-hopping and mixer exposure.

5.1(f) Case management and investigation

The AML solution shall, at a minimum, support:

Case Management and Investigation

What an examiner will look for
Listed as a minimum function in the Guidance Note (section 5), which also names “structured investigation workflows” as part of governance (section 4(b)). Detail under 5.7.
What this can look like
Alerts become cases that are assigned, investigated and decided in one workflow, with the decision recorded.

5.1(g) Regulatory and internal reporting

The AML solution shall, at a minimum, support:

Regulatory and Internal Reporting

What an examiner will look for
Listed as a minimum function in the Guidance Note (section 5); the CBN assesses the solution’s effectiveness “in supporting ... regulatory reporting” (Guidance Note, section 7). Detail under 5.8.
What this can look like
Regulatory returns and internal management reports are produced from the same data and cases the team works on.

5.1(h) Audit and governance, with logs and configuration trails

The AML solution shall, at a minimum, support:

Audit and Governance (including logs and configuration trails)

What an examiner will look for
The Guidance Note lists “audit, governance, and control logging” as a minimum function (section 5) and “clear audit trails” and “traceability of actions” as part of defensibility (section 4(a)). Detail under 5.9.
What this can look like
Every user action and every change to rules, thresholds and settings is logged and can be retrieved.

5.1(i) Data protection and security controls

The AML solution shall, at a minimum, support:

Data Protection and Security Controls relevant to AML/CFT/CPF activities

What an examiner will look for
Listed as a minimum function in the Guidance Note (section 5). Detail under 5.11.
What this can look like
Access, encryption and data-protection controls cover the AML data and the AML system.

5.1 Fraud kept apart from AML on a shared platform

The AML solution shall:

Where the same platform is also used to monitor fraud, its fraud-related capabilities shall be clearly segregated, governed, and tuned so as not to dilute AML/CFT/CPF detection effectiveness.

What an examiner will look for
A list of rules showing which are fraud and which are AML, with separate ownership and tuning records, and AML alert metrics tracked on their own so that a change to fraud rules can be seen not to reduce AML detection.
What this can look like
Fraud and AML rules are labelled and governed separately, and tuned by their own owners, even where their alerts share one customer view.

5.1, institution (i) A solution that fits the institution

The institution shall:

ensure that the AML Solution is commensurate with the nature, size, complexity, customer base and risk profile of its operations.

What an examiner will look for
The CBN assesses “the institution’s ability to demonstrate that its AML solution is appropriately configured to its risk profile” (Guidance Note, section 9), with sophistication “proportionate to their risk profile, size, and complexity” (section 8). Expect a written rationale linking products, channels, customer base and volumes to what the solution covers.
What this can look like
A document that maps each product, channel and customer segment to the screening and monitoring that covers it, reviewed when the business changes.

5.1, institution (ii) Availability and disaster recovery

The institution shall:

ensure appropriate availability, resilience, and disaster recovery arrangements so that AML monitoring and screening are not degraded during system outages or migrations.

What an examiner will look for
A continuity and disaster recovery plan that includes the AML solution, results of recovery tests, and records of past outages or migrations showing that screening and monitoring continued or caught up.
What this can look like
Recovery tests include the AML system, and events received during an outage are screened and monitored once it ends.

5.1, institution (iii) Review and upgrade as the business changes

The institution shall:

review and where necessary, upgrade the AML Solution in line with changes in business model, product mix, delivery channels and risk exposure.

What an examiner will look for
Records of reviews of the solution when products, channels or risks changed, and of the upgrades that followed.
What this can look like
Launching a product or channel includes a check that screening and monitoring cover it, and gaps are closed before launch.

5.2 Customer Due Diligence (CDD), Know Your Customer (KYC) and Know Business (KYB)

8 requirements

5.2(a)(i) End-to-end CDD, EDD, KYC and KYB support

The AML solution shall:

enable end-to-end support for CDD, EDD, KYC, and KYB processes, including automated risk profiling, transaction behavioural pattern analysis, use of historical data and detection of ML/TF/PF typologies.

What an examiner will look for
The roadmap must describe how data flows “customer onboarding → risk profiling” and give “at least one example demonstrating end-to-end data flow” (Template, section 4.4). Expect a walk-through of one customer from onboarding to risk profile to monitoring.
What this can look like
One customer record holds the KYC or KYB data, EDD status and risk profile, and the customer’s transaction history is used to update the profile.

5.2(a)(ii) Periodic and event-driven reviews

The AML solution shall:

support risk based CDD and EDD, including periodic and event driven reviews based on customers’ risk categories and behaviour.

What an examiner will look for
Review frequencies set by risk category, a list of the events that trigger a review, and reports showing reviews done on time and overdue ones followed up.
What this can look like
Higher-risk customers are reviewed more often, and events such as a change of ownership or a serious alert start a review automatically.

5.2(a)(iii) KYC linked to monitoring

The AML solution shall:

allow continuous data synchronisation and linkage between KYC/KYB records, customer risk profiles and transactional data, so that monitoring scenarios and alerts are evaluated in the context of the customer’s known profile, expected activity and risk classification.

What an examiner will look for
The Guidance Note requires “a consolidated, real-time or near real-time view of customer risk across all relevant data sources” (section 5). For each data source the roadmap must say whether integration is automated or manual and whether it is real-time, near real-time or batch (Template, section 5.3).
What this can look like
Changes to KYC data reach the monitoring system automatically, and rules can read the customer’s declared profile and current risk rating.

5.2(a)(iv) KYC on the alert screen

The AML solution shall:

ensure that, for any alert or case, investigators can view relevant KYC/KYB information (e.g., occupation, source of funds, income range, business activity, geography) alongside transactional history and prior case outcomes within a single interface.

What an examiner will look for
A demonstration of an alert or case screen showing the customer’s KYC or KYB data, transaction history and past case outcomes together, in line with the consolidated view the Guidance Note requires (section 5) and section 4(l) of the standards.
What this can look like
The investigator opens a case and sees occupation, source of funds, income range, business activity and geography next to the transactions and earlier case decisions, without switching systems.

5.2(b)(i) Automated onboarding with BVN and NIN checks

The institution shall:

implement, to the extent feasible, automated or semi-automated onboarding with customer identification processes supported by real-time data checks, in line with CBN KYC requirements. Where available, institutions should integrate with national identity databases, including the Bank Verification Number (BVN) and National Identification Number (NIN) systems, as infrastructure that supports real-time identity corroboration; i.e., confirming that presented identity credentials are linked to a registered record. Institutions should note that such database checks are one component of a broader identification and verification framework and do not, in isolation, constitute complete customer verification.

What an examiner will look for
“customer/KYC/BVN/NIN data” is a minimum data source the roadmap must integrate (Template, section 5.2), and “reliance on purely manual processes that do not ensure consistency, auditability, and timely detection will not meet regulatory expectations” (Guidance Note, section 8). Expect the onboarding flow, its BVN and NIN checks, and the other verification steps that sit beside them.
What this can look like
Onboarding checks the BVN or NIN in real time and also verifies the person by other means, and both results are recorded.

5.2(b)(ii) Tell the CBN which AML systems are in use

The institution shall:

Institutions shall notify the relevant CBN supervisory department of all AML Solutions in use, specifying the primary and supporting functional roles of each system. This notification shall be made at initial deployment, upon material change, and upon request by the CBN. This requirement supports the CBN’s maintenance of a sector-wide inventory of AML solution infrastructure.

What an examiner will look for
Copies of the notifications, listing each system and its primary and supporting roles, consistent with the “core systems supporting AML/CFT/CPF activities” described in the roadmap (Template, section 4.1).
What this can look like
An inventory of AML systems and what each does, kept current and sent to the supervisory department at deployment and on every material change.

5.2(b)(iii) Beneficial ownership

The institution shall:

exercise due diligence to understand and record complex ownership and control structures and beneficial ownership, including manual escalation where system capabilities are insufficient.

What an examiner will look for
Recorded ownership and control structures for corporate customers, the beneficial owners identified, and records of escalation where the system could not resolve the structure.
What this can look like
Ownership charts are kept for business customers, and structures the system cannot resolve go to a person, whose conclusion is recorded.

5.2(b)(iv) KYC data quality

The institution shall:

Ensure that CDD/KYC/KYB data used by the AML Solution is accurate, complete and kept up to date through periodic reviews and data quality controls.

What an examiner will look for
The roadmap must explain how the institution ensures “completeness of data”, “accuracy of data” and “consistency across systems” (Template, section 5.4). Expect data-quality reports and the remediation that followed.
What this can look like
Regular reports of missing or out-of-date KYC fields, with owners and deadlines to fix them.

5.3 Sanction Lists & PEP Screening

8 with rules 11 requirements

5.3(a)(i) Screening against sanctions lists and watchlists

The AML solution shall:

Integrate with relevant domestic and global sanctions, watchlists and other prescribed lists for screening of customers, beneficial owners, related parties, and transactions.

What an examiner will look for
Sanctions and watchlist data are a minimum data source, and “screening engines” a minimum system, in the roadmap’s integration plan (Template, sections 5.2 and 4.3). Expect the list of sources loaded and evidence that customers, owners, related parties and transactions are each screened.
What this can look like
Customers, beneficial owners, directors and payment counterparties are screened against the UN list, the Nigeria Sanctions List and the other lists that apply.
Rules that help
Sanctions screening

5.3(a)(ii) Matching that catches name variations

The AML solution shall:

include matching logic capable of detecting name variations and similarities, which may include AI- or fuzzy-matching techniques, provided these are subject to transparent configuration and validation.

What an examiner will look for
The documented matching configuration (method and thresholds), test results showing that misspellings, transliterations and reordered names are caught, and a record of changes to the configuration.
What this can look like
A fuzzy-matching threshold chosen by testing against a set of name variants, written down, and re-tested when it changes.

5.3(a)(iii) Lists kept up to date

The AML solution shall:

support real-time or near real-time updates of sanctions and watchlists and ensure that new or updated entries are promptly reflected in screening.

What an examiner will look for
Logs showing when each list update was published and when it was loaded, and that existing customers were rescreened against the new entries.
What this can look like
Lists refresh automatically, and when a name is added the whole customer base is rescreened against it.

5.3(a)(iv) Logs that prove screening works

The AML solution shall:

Possess the capability to generate comprehensive logs that provide verifiable evidence of timely updates and the effectiveness of screening processes.

What an examiner will look for
The logs the clause describes: list versions and load times, screening runs, and the matches each produced and how they were resolved.
What this can look like
An update log with each list version, when it was loaded, and how many customers were rescreened, next to a record of matches and their outcomes.

5.3(a)(v) Internal watchlists

The AML solution shall:

Permit inclusion and management of institution-specific internal watchlists.

What an examiner will look for
The internal list, who added each entry and why, and screening matches against it.
What this can look like
The institution’s own list of confirmed bad actors (names, accounts, phones, devices) is screened like an external list, with an owner and a reason for every entry.

5.3(a)(vi) PEP and high-risk lists

The AML solution shall:

Integrate PEP and high-risk customer lists/registers and automatically flag PEPs and other high-risk individuals and entities.

What an examiner will look for
PEP databases are a minimum data source in the roadmap (Template, section 5.2). Expect the PEP data used, the customers flagged, and the enhanced due diligence applied to them.
What this can look like
PEP data is loaded into screening, and a PEP match flags the customer for enhanced due diligence automatically.

5.3(a)(vii) Adverse media

The AML solution shall:

Support adverse media or negative news monitoring.

What an examiner will look for
The glossary defines adverse media monitoring as “Continuous screening of news and open-source content to identify negative information related to customers, beneficial owners, or related parties that may indicate ML/TF/PF risk.” Expect the news and open sources used, evidence that customers, beneficial owners and related parties are screened on an ongoing basis, and the matches reviewed and their outcomes.
What this can look like
Customers, beneficial owners and related parties are screened against news and open sources continuously, not only at onboarding, and matches are reviewed like other screening matches.

5.3(a)(viii) Stop transactions and onboarding on a confirmed sanctions match

The AML solution shall:

Be capable of automated interdiction or hold on transactions, or blocking of onboarding, where there is a confirmed sanctions match, in line with CBN and legal requirements.

What an examiner will look for
Configuration showing that a confirmed sanctions match holds the transaction or blocks onboarding, and records of transactions held and onboardings blocked. The glossary defines real-time screening as evaluating a customer or transaction “before the transaction is authorised or onboarding process is considered completed.”
What this can look like
A payment to a confirmed match is stopped before it completes, and onboarding cannot finish while a match is open.
Rules that help
Sanctions screening

5.3(b)(i) Screening at onboarding and ongoing

The institution shall:

Ensure appropriate, risk-based screening at onboarding, periodically and on a continuous basis, in line with extant regulations.

What an examiner will look for
A screening policy that sets when screening happens, and records showing screening at onboarding, at periodic reviews and on list changes.
What this can look like
Customers are screened at signup, again at each periodic review, and whenever the lists change.

5.3(b)(ii) Procedure for resolving matches

The institution shall:

maintain documented procedures for triaging, reviewing, and resolving potential matches, including escalation and reporting.

What an examiner will look for
The written procedure and cases showing it was followed. The Guidance Note names “structured investigation workflows” as part of governance (section 4(b)).
What this can look like
A written procedure that says who reviews a match, which identifiers clear it, when it goes to the compliance officer, and when it is reported.

5.3(b)(iii) Show that screening works

The institution shall:

be able to demonstrate, upon request, that sanctions and PEP screening processes are effective, including evidence of resolved alerts, decisions and actions taken.

What an examiner will look for
The clause names the evidence: resolved alerts, the decision on each and the action taken. The Guidance Note expects “measurable outcomes (including false positive management)” (section 4(c)).
What this can look like
For any period, a list of screening matches with how each was resolved, by whom, why, and what was done.

5.4 Risk Assessment

8 requirements

5.4(a)(i) Configured to the institution’s risk assessment

The AML solution shall:

be configurable to reflect the financial institution’s documented risk appetite and risk assessment, including rule configurations, risk scenarios, and alert thresholds.

What an examiner will look for
The CBN assesses “the institution’s ability to demonstrate that its AML solution is appropriately configured to its risk profile” (Guidance Note, section 9). Expect each rule, scenario and threshold to trace to the documented risk assessment.
What this can look like
Rules and thresholds can be changed by the compliance team, and each one is linked to the risk it addresses in the risk assessment.

5.4(a)(ii) Dynamic customer risk rating

The AML solution shall:

Support automated risk assessment at onboarding, dynamically adjusting customer risk profile in response to new data, behavioural changes and external risk factors.

What an examiner will look for
The roadmap must show the flow “customer onboarding → risk profiling” (Template, section 4.4). Expect the risk model, a customer’s rating at onboarding, and examples of ratings that changed after new data or behaviour.
What this can look like
A customer’s risk rating is set at onboarding and moves automatically when their data, alerts or behaviour change.

5.4(a)(iii) Enterprise-wide risk view

The AML solution shall:

support enterprise-level ML/TF/PF risk identification, measurement, and assessment (for example, by aggregating data across products, segments, and channels).

What an examiner will look for
Enterprise-level risk reports produced from the solution’s data, by product, segment and channel, used in the institution’s risk assessment.
What this can look like
Reports of risk exposure by product, segment and channel feed the annual enterprise risk assessment.

5.4(a)(iv) Governed AI and machine learning

The AML solution shall:

Where AI/ML-based models are used, support adaptive learning and model calibration under a documented governance framework, including human oversight and explainability.

What an examiner will look for
Applies only where AI or machine learning is used; the standards do not require it (Guidance Note, section 2). The Guidance Note expects “explainable decisions” and “model validation and change control” (section 4). Expect the governance framework, records of approved recalibrations, and the explanation an analyst sees for a score.
What this can look like
Model changes are approved before use, and analysts see the main reasons behind each score.

5.4(a)(v) Report risk rating changes and why

The AML solution shall:

generate reports on changes in customer risk classification and the drivers of such changes.

What an examiner will look for
A report listing customers whose risk classification changed in a period, with the factor that caused each change.
What this can look like
A monthly list of upgrades and downgrades in customer risk, each with its reason.

5.4(a)(vi) External data in risk scoring

The AML solution shall:

where applicable, incorporate and analyse credible external data sources to enhance risk scoring and identification of emerging threats.

What an examiner will look for
External data sources are a minimum data source “where applicable” in the roadmap (Template, section 5.2). Expect a list of the external sources used and how each affects risk scores.
What this can look like
Country risk lists, adverse media or registry data are used as factors in customer risk scores.

5.4(b)(i) Risk assessment drives configuration

The institution shall:

conduct and document periodic ML/TF/PF risk assessments at enterprise and business-line level and ensure that the configuration of the AML Solution reflects these assessments.

What an examiner will look for
The enterprise and business-line risk assessments, and a mapping from their risks to the solution’s configuration, which the CBN checks when it assesses whether the solution “is appropriately configured to its risk profile” (Guidance Note, section 9).
What this can look like
The risk assessment has a table linking each significant risk to the scenarios and thresholds that monitor it.

5.4(b)(ii) Keep risk assessments and the changes they caused

The institution shall:

generate and retain risk assessment reports, together with evidence of resulting changes to scenarios, thresholds, and control activities.

What an examiner will look for
The retained risk assessment reports and dated records of the scenario, threshold and control changes made because of them.
What this can look like
After each assessment, a change log shows which scenarios and thresholds were changed and which finding each change answers.

5.5 Transaction Monitoring & Risk-Based Analyses

4 with rules 13 requirements

5.5(a)(i) Risk-based transaction monitoring

The AML solution shall:

support risk-based transaction and activity monitoring using appropriate techniques (including, where suitable, predictive analytics, anomaly detection, behavioural pattern recognition, and automated risk scoring) to detect potential ML/TF/PF activity.

What an examiner will look for
Rules, models or both are acceptable (Guidance Note, section 2); what counts is “credible detection capability” (section 4(c)). The roadmap must explain how the solution will “improve detection of ML/TF/PF risks” and “enhance risk-based monitoring” (Template, section 7.3).
What this can look like
Rules for known typologies, plus checks that compare a customer’s activity with their own history, with stricter settings for higher-risk customers. An institution that offers virtual asset services also runs scenarios on its virtual asset transfers, such as chain-hopping and mixer exposure.

5.5(a)(ii) Pre-emptive alerts and temporary holds

The AML solution shall:

have the capability, where justified by risk, to generate pre-emptive alerts for high-risk scenarios that allow institutions to take preventive or mitigating action, such as requiring enhanced verification, or placing a temporary hold on a transaction, pending review.

What an examiner will look for
The high-risk scenarios configured to act before a transaction completes, and records of holds or verification requests and how each was resolved. The procedures behind them are under 5.5(b)(iv).
What this can look like
A high-risk pattern holds the payment and opens a case, and an analyst releases or rejects it.

5.5(a)(iii) Multiple scenarios, configurable rules and segmentation

The AML solution shall:

incorporate comprehensive transaction monitoring using multiple risk scenarios, configurable rules and customer segmentation, aligned to documented risk assessments and typologies.

What an examiner will look for
The list of scenarios in use, each linked to a typology or risk in the risk assessment, with the customer segments it covers. The CBN looks at whether the solution “is appropriately configured to its risk profile” (Guidance Note, section 9).
What this can look like
A scenario library covering the institution’s main typologies, with thresholds that differ by segment (for example individuals and businesses). An institution that offers virtual asset services also runs scenarios on its virtual asset transfers, such as chain-hopping and mixer exposure.

5.5(a)(iv) Monitoring uses KYC attributes

The AML solution shall:

ensure that monitoring scenarios and models make risk-based use of relevant CDD/KYC/KYB attributes (such as occupation, declared income or turnover range, customer segment, product profile, geography, and delivery channel) in addition to raw transactional patterns.

What an examiner will look for
Scenarios whose conditions use KYC or KYB attributes, such as declared turnover or segment. Section 4 of the standards says solutions “without effective linkage to CDD/KYC/KYB information and customer risk assessments will not be regarded as compliant”.
What this can look like
A rule compares a customer’s monthly total with their declared income or turnover, and thresholds differ by segment, product and channel.

5.5(a)(v) Related parties, networks and peer groups

The AML solution shall:

Support related-party mapping, network analysis and peer-grouping to improve identification of unusual or suspicious patterns. Where AI/ML is used, provide clear, auditable information on the key factors that contributed to the alerts to support human review and explainability (like unusual transaction patterns or behaviour compared to similar customers).

What an examiner will look for
A demonstration of related-party and network views and of peer-group comparisons used in monitoring. Where AI is used, the Guidance Note expects “explainable decisions” (section 4(a)): the alert shows the factors behind it.
What this can look like
Accounts that share identifiers or exchange money are drawn as a network, and a customer’s activity can be compared with similar customers.

5.5(a)(vi) One view of the customer on every alert

The AML solution shall:

Present, for each alert, a consolidated view of the customer’s key KYC/KYB data, risk score, recent transactional behaviour and any previous alerts or cases, to support informed human decision-making on whether to close, escalate or report.

What an examiner will look for
A demonstration of the alert screen showing KYC or KYB data, risk score, recent transactions and previous alerts and cases together: the consolidated view the Guidance Note requires (section 5) and section 4(l) of the standards describes.
What this can look like
Opening an alert shows the customer’s profile, current risk score, recent transactions and past alerts and case outcomes on one screen.

5.5(b)(i) Independent validation of AI and ML models

The institution shall:

Perform independent validation of all artificial intelligence and machine learning models at least annually and upon significant change covering accuracy, performance drift, fairness audits, bias testing, and human review where relevant and appropriate to the institution’s risk profile. For the purposes of this requirement, “independent” is defined in the Glossary.

What an examiner will look for
Dated validation reports, at least yearly and after significant changes, covering accuracy, drift, fairness and bias, by a validator who meets the glossary’s definition: “organisationally separate from, and has no operational responsibility for, the model or system being validated”, for AI/ML models “a team or individual with no involvement in model development or day-to-day configuration”, and possessing “the requisite technical competence to assess model accuracy, performance drift, fairness, and bias”. Model validation is part of governance in the Guidance Note (section 4(b)) and a minimum governance element in the roadmap (Template, section 6.2).
What this can look like
Each year and after each significant change, a validator who took no part in building the model or in configuring it day to day, sits outside the team that runs it and has the skills to test it (an internal model-risk or internal audit function, or an outside party) validates it and records the findings and fixes.

5.5(b)(ii) False positive and false negative thresholds

The institution shall:

Define, document, and periodically review false positive and false negative thresholds appropriate to their risk profile, product mix, customer base, and transaction volumes.

What an examiner will look for
Documented acceptable false positive and false negative levels and the reviews of them. False positive rate is a minimum metric in the roadmap, with current and target values (Template, section 7.2), and the Guidance Note expects “false positive management” (section 4(c)).
What this can look like
For each scenario, a documented acceptable false positive rate, and a method such as below-threshold testing to estimate what it misses.

5.5(b)(iii) Document and justify thresholds and tuning

The institution shall:

Document and justify threshold settings, scenario parameters and tuning decisions at least annually, and whenever material changes are made.

What an examiner will look for
A tuning record for each scenario with the analysis behind its thresholds, dated at least yearly and at each material change. The Guidance Note names “change control” as part of governance (section 4(b)); change management and approval is a minimum governance element (Template, section 6.2).
What this can look like
Each scenario has a file with its current thresholds, the data analysis that justified them, and every change with its date and approver.

5.5(b)(iv) Procedures for holds and interventions

The institution shall:

Develop and document procedures for pre-emptive interventions arising from high-risk patterns, ensuring such measures are proportionate and consistent with customer rights and experience.

What an examiner will look for
The written procedure for holds and other interventions (who can apply and lift them, how long they last, how the customer is told) and records showing it was followed.
What this can look like
A hold procedure with a time limit, a named reviewer and customer contact steps.

5.5(b)(v) Governance of automated actions and logic changes

The institution shall:

Establish governance arrangements (e.g., model committee, change-control process) overseeing authorisation and periodic review of automated actions and major changes to monitoring logic.

What an examiner will look for
Terms of reference of the committee or change process, minutes approving automated actions and major logic changes, and periodic reviews. The CBN assesses “governance arrangements, including ownership, oversight, model validation, and change management” (Guidance Note, section 7).
What this can look like
A committee approves new automated actions and major rule changes before they go live and reviews them on a schedule.

5.5(b)(vi) Time limit for reviewing high-risk alerts

The institution shall:

Institutions shall define and adhere to internal service level standards specifying the maximum timeframe within which high-risk alerts shall be reviewed and a disposition decision recorded. These standards shall be documented and approved by senior management.

What an examiner will look for
The approved service level standard and ageing reports showing high-risk alerts decided within it. The Guidance Note expects “timely investigation and resolution” (section 4(c)); investigation turnaround time is a minimum metric in the roadmap (Template, section 7.2).
What this can look like
A senior-management-approved maximum time for deciding high-risk alerts, and a weekly report of alerts that are close to or past it.

5.5(b)(vii) Conditions for automated alert closure

The institution shall:

Where the institution permits automated alert closure, ensure that:

  • such automation is limited to clearly low-risk scenarios defined in the AML governance framework. For the purposes of these Baseline Standards, a “clearly low-risk scenario” eligible for automated closure is one that meets all of the following criteria: (a) the alert is generated by a rule or threshold reviewed, approved and documented by the institution’s AML governance committee; (b) the closure decision relies explicitly on both transactional context and current KYC/KYB data; (c) the customer’s risk classification is Low and has not changed within the preceding review period; and (d) no prior unresolved alerts or open cases exist for the same customer at the time of closure;
  • the decision rules explicitly rely on both transactional and KYC/KYB context;
  • periodic back-testing confirms that automated closures do not mask suspicious activity;
  • Institutions that permit automated alert closure shall notify the CBN of this practice, the categories of scenarios subject to automation, and the institution-defined threshold for the proportion of total alerts eligible for automated closure. This notification shall be made upon adoption and following any material change. The institution’s defined threshold shall be documented, approved by the AML governance committee, and reviewed at least annually. Any quarter in which automated closures materially exceed the institution’s documented threshold shall be escalated promptly to senior management and reported to the CBN; and
  • Governance arrangements for automated alert closure shall be subject to independent review by internal audit at least annually, separate from the compliance.
What an examiner will look for
Only applies if alerts are closed automatically. The clause names the evidence: the committee-approved list of eligible scenarios meeting all four criteria, the closure logic, back-testing results, the notification to the CBN, the documented threshold and its annual review, quarterly figures against it, and internal audit’s annual review.
What this can look like
Either no automated closure, or a short list of approved low-risk scenarios whose closures check the customer’s current Low rating and open cases, with quarterly counts and back-testing.

5.6 Fraud Monitoring and Detection

2 with rules 8 requirements

The section opens: Where the AML Solution is also used for fraud monitoring, the following standards apply in addition to AML/CFT/CPF expectations:

Section 4(j) separately requires every institution to run automated fraud monitoring, on the AML platform or on a separate system.

5.6(a)(i) Real-time fraud monitoring across channels

The AML solution shall:

Monitor activities across relevant channels (e.g. cards, e-channels, deposits, lending) in real time or near real time to detect unusual patterns indicative of fraud. For the purposes of this section, “real time” and “near real time” shall have the meanings set out in the Glossary, which distinguish between the standards applicable to screening and to transaction monitoring respectively. Fraud monitoring controls, particularly for card and electronic channels, shall be calibrated to the faster intervention timelines appropriate to those channels.

What an examiner will look for
Evidence that each channel’s events reach fraud monitoring and how fast, with interventions that act before completion on card and electronic channels. The glossary defines real-time monitoring as able to “trigger a pre-emptive action within stipulated timeframe that can prevent the completion of a suspicious activity (e.g., before funds are irreversibly settled)”. Transaction channels (mobile, USSD, cards, POS) are minimum data sources, each with its frequency stated (Template, sections 5.2 and 5.3).
What this can look like
Card, mobile, USSD and transfer events are checked as they happen, and the riskiest are held before the money moves.

5.6(a)(ii) Update fraud rules as fraud changes

The AML solution shall:

Enable updating of fraud-related risk rules and models based on observed fraud trends and incidents across customers, accounts, and channels.

What an examiner will look for
Change records showing fraud rules or models updated after fraud incidents or trends, with the incident each change responds to.
What this can look like
After a new fraud pattern is confirmed, the fraud team changes or adds a rule, tests it on past data and puts it live.

5.6(a)(iii) One workflow for AML and fraud, with separate teams

The AML solution shall:

Support a unified workflow or tightly integrated workflows for managing alerts and investigations across AML, CFT, CPF and fraud with clear segregation of responsibilities to applicable units.

What an examiner will look for
A demonstration of the case workflow handling AML and fraud alerts, and the queues, roles or permissions that keep each team to its own work.
What this can look like
Fraud and AML alerts share one customer view, and go to separate queues worked by separate teams.

5.6(a)(iv) Fraud registries and blacklists

The AML solution shall:

Interface with relevant fraud registries, internal blacklists and external databases for cross-verification and screening.

What an examiner will look for
The registries, internal blacklists and external databases connected, and matches found against them.
What this can look like
Customers and counterparties are checked against the institution’s own blacklist and industry watch-lists, and a match raises an alert.

5.6(a)(v) Fraud trends from history

The AML solution shall:

Analyse historical data to identify fraud trends and generate predictive insights, where appropriate.

What an examiner will look for
Fraud trend reports built from historical data, and the rule or model changes they led to.
What this can look like
A periodic fraud trend report by channel and type, used to decide which rules to add or change.

5.6(a)(vi) Traceability of fraud events

The AML solution shall:

Maintain full traceability of fraud-related events, including decisions and outcomes.

What an examiner will look for
For a sample of fraud events, the trail from event to alert to decision to outcome. The Guidance Note expects “traceability of actions” (section 4(a)).
What this can look like
Any fraud alert can be followed from the transaction that raised it to the decision and the final outcome.

5.6(b)(i) A unified financial crime architecture where risk justifies it

The institution shall:

where the risk justifies it, having regard to the institution’s transaction volumes, product complexity, customer base and documented risk assessment, implement a unified financial crime risk architecture in which AML and fraud systems share a common data lake, analytics models, and coordinated case management, while ensuring that AML/CFT/CPF requirements remain fully met and not subordinated to fraud priorities. Institutions classified as High or Above Average risk within their respective subsectors, or those with material volumes of electronic or card-based transactions, shall be expected to demonstrate a credible roadmap towards such integration.

What an examiner will look for
For institutions in scope, a roadmap towards shared data, models and case management. The Implementation Roadmap Template asks for current and target architecture diagrams and how the target will “eliminate fragmentation” (Template, sections 4.1, 4.2 and 4.6), and has a fraud monitoring and detection row in its core table (section 3).
What this can look like
A target architecture where fraud and AML read the same customer and transaction data and work cases in coordinated queues, with dated milestones.

5.6(b)(ii) Fraud indicators feed AML risk

The institution shall:

Ensure that material fraud indicators are incorporated as risk factors into the customer’s overall ML/TF/PF risk profile and relevant transaction monitoring scenarios, where appropriate.

What an examiner will look for
The customer risk model showing fraud indicators as factors, and AML scenarios that use them.
What this can look like
A confirmed fraud case raises the customer’s AML risk rating, and AML rules can use fraud flags as conditions.

5.7 Case Management

6 requirements

5.7(a)(i) Cases created, assigned and tracked automatically

The AML solution shall:

Provide Enterprise Case Management (ECM) capability that automates the creation, assignment, prioritisation and tracking of cases arising from alerts (AML/CFT/CPF and where relevant, fraud).

What an examiner will look for
A demonstration of alerts turning into cases, being assigned and prioritised, and tracked to closure. The Guidance Note names “structured investigation workflows” as part of governance (section 4(b)).
What this can look like
Each alert opens or joins a case automatically, the case goes to the right queue and person, and its status and deadline are tracked.

5.7(a)(ii) Role-based workflows, maker-checker and escalation

The AML solution shall:

Enable role-based workflows, including Maker-Checker functionality and escalation paths to ensure robust case review and resolution.

What an examiner will look for
A demonstration of a case needing a second person to approve a decision, and of escalation paths by role. The glossary defines maker-checker as a “Role-based control that requires more than one user to complete a workflow / process.” Structured investigation workflows are part of governance (Guidance Note, section 4(b)).
What this can look like
An analyst proposes a decision, a reviewer approves it, and cases can be escalated to the compliance officer.

5.7(a)(iii) Audit trail on cases

The AML solution shall:

Maintain full audit trails of all actions taken on cases, including timestamps, users, decisions and supporting rationale.

What an examiner will look for
The full history of a sample of cases: every action, with who did it, when, the decision and the reason. The Guidance Note expects “clear audit trails” and “traceability of actions” (section 4(a)).
What this can look like
Each case shows a timeline of every action, user and decision, with the reason the analyst recorded.

5.7(a)(iv) Case reports for management and supervisors

The AML solution shall:

Support the generation of reports on case volumes, ageing, outcomes, and key trends to inform management oversight, support compliance testing, operational review, and supervisory inspections.

What an examiner will look for
Case reports by volume, age, outcome and trend, produced from the system. Alert volumes, STR conversion rate and investigation turnaround time are minimum metrics in the roadmap (Template, section 7.2).
What this can look like
A monthly case report: cases opened and closed, ageing buckets, outcomes, and how many led to a report.

5.7(b)(i) Prompt, documented investigations

The institution shall:

Ensure that cases are investigated promptly and adequately document outcomes and rationale.

What an examiner will look for
Case turnaround figures and a sample of closed cases with their documented outcome and reasoning. The Guidance Note expects “timely investigation and resolution” (section 4(c)); investigation turnaround time is a minimum metric (Template, section 7.2).
What this can look like
Cases are closed within set times, and each closure states what was found and why it was closed or escalated.

5.7(b)(ii) Learn from closed cases

The institution shall:

Periodically review and analyse closed cases to identify emerging patterns, control weaknesses and opportunities for scenario tuning and escalate material findings to Senior Management and, where necessary, to regulators.

What an examiner will look for
Periodic reviews of closed cases, the tuning changes and control fixes they led to, and the escalations to senior management.
What this can look like
A quarterly look-back at closed cases that leads to rule changes and is reported to senior management.

5.8 Reporting

5 requirements

5.8(a)(i) Automated or semi-automated regulatory reports

The AML solution shall:

automated or semi-automated generation of regulatory reports such as Suspicious Transaction Report (STR), Suspicious Activity Reports (SARs), Currency Transaction Report (CTR), Foreign Currency Transaction Report (FTR) and other applicable AML/CFT/CPF returns, with configurable formats and schedules in line with CBN reporting instructions.

What an examiner will look for
Reports generated from cases and data in the required formats, and the roadmap’s end-to-end flow “alert → case → regulatory reporting” (Template, section 4.4). The CBN assesses effectiveness “in supporting detection, investigation, and regulatory reporting” (Guidance Note, section 7).
What this can look like
An STR or CTR is drafted from the case data in the required format, reviewed by the compliance team and filed.

5.8(a)(ii) Management information to compliance, management and the Board

The AML solution shall:

Periodic internal management information (MI) reporting to, at a minimum, the Chief Compliance Officer, senior management, the Executive Compliance Officer, and the Board, to support oversight of AML/CFT/CPF risks and the effectiveness of transaction monitoring.

What an examiner will look for
MI reports with their dates and recipients, covering risk and monitoring effectiveness, for example the metrics the roadmap asks institutions to track: alert volumes, false positive rate, STR conversion rate and investigation turnaround time (Template, section 7.2).
What this can look like
A monthly MI pack from the system to the Chief Compliance Officer, the Executive Compliance Officer and senior management, and a quarterly one to the Board.

5.8(a)(iii) External reports only to lawful authorities

The AML solution shall:

External reporting only to regulators and other authorities with a lawful mandate, ensuring the confidentiality and appropriateness of all shared information.

What an examiner will look for
Controls limiting who can export or send reports and to whom, and logs of external reports sent.
What this can look like
Only named roles can file or export reports, and every report sent outside is logged with its recipient.

5.8(b)(i) Timely, accurate, complete filings

The institution shall:

Ensure timely, accurate and complete submission of reports to the CBN, NFIU and other competent authorities in the required formats and channels.

What an examiner will look for
Submission records and acknowledgements, matched against filing deadlines and the cases they came from. The CBN assesses the solution’s effectiveness “in supporting ... regulatory reporting” (Guidance Note, section 7).
What this can look like
A register of every report filed, with its due date, filing date and acknowledgement.

5.8(b)(ii) Review and approval of regulatory reports

The institution shall:

Establish internal governance for the review and approval of all regulatory reports, ensuring they are consistent with underlying data, case-management outcomes, and the institution’s documented investigative rationale.

What an examiner will look for
The approval procedure, approver sign-offs on reports, and a sample of reports traced to their case and data.
What this can look like
Every report is reviewed and approved by the compliance officer before filing, and each one links to the case it came from.

5.9 Audit and Governance

11 requirements

5.9(a)(i) Tamper-proof audit trail of all activity

The AML solution shall:

maintain a comprehensive, tamper-proof and immutable audit trail of all system and user activities, including configuration changes, access events, alert dispositions, and report generation that can be retrieved without disrupting ongoing operations.

What an examiner will look for
The audit trail itself, covering configuration changes, access events, alert decisions and report generation, and evidence that entries cannot be edited or deleted. The Guidance Note expects “clear audit trails” (section 4(a)); the glossary defines the audit trail as a “Tamper-proof and immutable, comprehensive log”.
What this can look like
An append-only log of every login, setting change, alert decision and report, that users cannot alter.

5.9(a)(ii) Who, when and what for every action

The AML solution shall:

record and retain key information such as identity of the user, date, timestamp, and nature of activity performed within the system.

What an examiner will look for
Sample log entries showing user, date and time, and the action taken. The Guidance Note expects “traceability of actions” (section 4(a)).
What this can look like
Every log entry carries the user, the exact time and what was done.

5.9(a)(iii) History kept for the legal retention period

The AML solution shall:

preserve a historical record of all system actions in line with extant laws and regulatory retention periods.

What an examiner will look for
The retention settings for system records, matched to the retention periods in law, and retrieval of an old record.
What this can look like
System history is kept for at least the legal retention period and can still be searched.

5.9(a)(iv) Search and trace end to end

The AML solution shall:

provide search and retrieval capabilities that enable authorised users to query, review and trace system and user activities end-to-end, to support review by Internal Audit, Compliance, Risk Management, and regulators.

What an examiner will look for
A live search tracing one activity end to end, by an authorised reviewer. The Guidance Note expects “traceability of actions” (section 4(a)).
What this can look like
An auditor can search by user, customer, alert or date and follow the trail from start to finish.

5.9(a)(v) Automated audit and governance reports

The AML solution shall:

generate automated audit and governance reports to support compliance, substantive testing, operational review, and supervisory inspections.

What an examiner will look for
Audit and governance reports generated by the system, such as user access, configuration change and alert disposition reports.
What this can look like
Standard reports of user access, changes to rules and settings, and alert decisions, available on demand for an inspection.

5.9(a)(vi) Links across customer, transaction, alert, action and report

The AML solution shall:

support forensic investigation by maintaining verifiable linkages across all relevant data elements including, at a minimum, customer data, transaction data, alerts, user actions, and regulatory returns.

What an examiner will look for
A trace from a regulatory return back to its case, alerts, transactions, customer and the users who acted. The roadmap must give “at least one example demonstrating end-to-end data flow” (Template, section 4.4).
What this can look like
From an STR, a reviewer can open the case, its alerts, the transactions behind them, the customer, and every user action along the way.

5.9(a)(vii) Retrieval without disrupting operations

The AML solution shall:

retrieve audit logs, workflow histories, and transaction trails without disrupting ongoing operations.

What an examiner will look for
A demonstration of retrieving logs and trails while monitoring and screening keep running.
What this can look like
Audit retrieval runs separately from live monitoring, so large exports do not slow screening.

5.9(b)(i) Documented governance framework

The institution shall:

establish a documented governance framework for the AML Solution, covering roles and responsibilities for ownership, configuration, change management, model validation, access rights and incident handling.

What an examiner will look for
The framework document. The Guidance Note expects “defined ownership and oversight” (section 4(b)), and the roadmap lists minimum governance elements: ownership, model validation, change management, oversight and reporting, escalation, documentation and audit trail (Template, section 6.2), with evidence available or “governance will be treated as not established” (section 6.3).
What this can look like
A governance document naming the system owner and who configures, changes, validates, grants access and handles incidents.

5.9(b)(ii) Segregated, reviewed access rights

The institution shall:

ensure that access rights, privileges and configuration authorities are appropriately segregated and periodically reviewed.

What an examiner will look for
The access matrix showing who can configure and who can decide, and records of periodic access reviews.
What this can look like
The people who change rules are not the only ones who approve them, and access is reviewed on a schedule.

5.9(b)(iii) Independent audit of the solution

The institution shall:

subject the AML Solution and its governance framework to periodic review by independent internal audit and, where appropriate, external assurance providers.

What an examiner will look for
Internal audit reports on the AML solution and its governance, and external assurance reports where used.
What this can look like
Internal audit includes the AML system in its plan and reports findings that are tracked to closure.

5.9(b)(iv) Documented training

The institution shall:

ensure that AML/CFT/CPF teams responsible for the operation, configuration, and oversight of the AML Solution receive regular training on system usage, emerging ML/TF/PF typologies, and relevant regulatory developments. Training activities shall be documented and records maintained for supervisory inspection.

What an examiner will look for
The clause names the evidence: training records for the teams that run, configure and oversee the solution.
What this can look like
A training log covering system use, new typologies and regulatory changes, with attendance for each team member.

5.10 System Integration & Scalability

11 requirements

5.10(a)(i) Real-time integration with core banking and KYC

The AML solution shall:

support secure, bidirectional integration with core banking, customer information/KYC systems and other relevant applications to ensure real-time or near real-time exchange of customer, account and transaction data for monitoring, screening and case management.

What an examiner will look for
The integration plan with each source system, method and frequency (Template, sections 4.3 and 5), and evidence such as “system interfaces, data flows, or sample outputs”. The Guidance Note says implementations relying on “fragmented or partially integrated systems without robust, automated integration will not meet regulatory expectations” (section 5).
What this can look like
Customer, account and transaction events flow into the AML system as they happen, and decisions flow back to core systems.

5.10(a)(ii) Real-time processes not degraded

The AML solution shall:

ensure that real-time monitoring and screening processes are not unduly interrupted, impacted or degraded by other system tasks or performance constraints.

What an examiner will look for
Performance and latency monitoring for real-time screening and monitoring, including during peak load and batch jobs.
What this can look like
Real-time checks keep their response time during batch jobs and peaks, and slowdowns raise an alert.

5.10(a)(iii) Documented, standards-based APIs

The AML solution shall:

provide well-documented, standards-based Application Programming Interfaces (APIs) or interfaces for integration with internal and where applicable, external systems.

What an examiner will look for
The API documentation and the interfaces in use. The roadmap asks for “Supporting evidence of integration (e.g. system interfaces, data flows, or sample outputs)” (Template, section 4.3).
What this can look like
Published API documentation that the institution’s engineers use to connect their systems.

5.10(a)(iv) Standard API data format

The AML solution shall:

standardise the format for exchange of data through the API in line with extant regulations.

What an examiner will look for
The data formats used on the API and the regulations they follow.
What this can look like
A fixed, documented schema for the data exchanged through the API.

5.10(a)(v) Standard data exchange formats

The AML solution shall:

support consistent, standardised data exchange formats in line with relevant CBN and national requirements.

What an examiner will look for
The exchange formats in use and the CBN or national requirements they follow.
What this can look like
Data exchanged with other systems and authorities uses the formats those requirements set.

5.10(a)(vi) Works with legacy and third-party systems

The AML solution shall:

offer flexibility to integrate with legacy systems and third-party services, if required.

What an examiner will look for
Integrations with legacy systems and third-party services in place, or the method available for them.
What this can look like
File-based or API integration for older core systems, alongside integrations with outside services.

5.10(a)(vii) Scales with volumes, products and channels

The AML solution shall:

be capable of handling increasing transaction volumes, new products, and channels without compromising performance or detection quality.

What an examiner will look for
Capacity tests or performance records at higher volumes, and how new products and channels were added.
What this can look like
Load tests above current peak volumes, and new products added as new event types without a new system.

5.10(b)(i) Scalable, with data integrity and secure transmission

The institution shall:

deploy AML Solutions that are scalable, maintain data integrity and ensure secure data transmission.

What an examiner will look for
Data integrity controls and encryption in transit. The roadmap must explain how the institution ensures completeness, accuracy and consistency of data across systems (Template, section 5.4).
What this can look like
Reconciliations between source systems and the AML system, and encrypted connections between them.

5.10(b)(ii) Testing and review of system changes

The institution shall:

subject major integration projects and system changes to appropriate testing, change-control, and post-implementation review, including evaluation of impacts on AML/CFT/CPF controls.

What an examiner will look for
Test plans and results, change approvals, and post-implementation reviews for major changes. The CBN assesses “change management” (Guidance Note, section 7); the roadmap timeline must include “testing and validation activities” (Template, section 9.2).
What this can look like
Every major change has a test plan, an approval and a review after go-live that checks alert volumes and screening still work.

5.10(c) Shared services only with CBN approval

The institution shall:

With prior approval of the CBN, institutions may operate AML Solutions under shared services or other centralised arrangements in line with the provisions of the Guidelines for Shared Services Arrangements for Banks and Other Financial Institutions or any other arrangement acceptable to the CBN. In such cases, responsibilities for compliance, data protection and system performance shall remain clearly allocated and documented.

What an examiner will look for
Where a shared or group arrangement is used: the CBN’s prior approval and the document allocating responsibilities. The Guidance Note says outsourcing “does not transfer” regulatory responsibility (section 6).
What this can look like
A group-wide AML system run under a CBN-approved shared services arrangement, with a written split of responsibilities.

5.10(d) No standalone transaction feeds for higher-risk institutions

The institution shall:

The CBN considers it unacceptable for institutions rated High or Above Average risk within their respective subsectors to operate AML solutions that rely solely on standalone transaction feeds. Such institutions are required to ensure that their AML systems are fully integrated with KYC and KYB repositories, as well as customer risk profiles.

What an examiner will look for
For institutions rated High or Above Average risk: integration of the AML system with KYC and KYB data and customer risk profiles, shown in the architecture and integration plan (Template, sections 4 and 5). Fragmented implementations “will not meet regulatory expectations” (Guidance Note, section 5).
What this can look like
The AML system receives KYC and KYB data and risk ratings as well as transactions, and rules use them.

5.11 Security & Data Protection

9 requirements

5.11(a)(i) Collect and store AML data lawfully

The AML solution shall:

collect and store data necessary for AML/CFT/CPF compliance in a manner consistent with regulatory requirements.

What an examiner will look for
A data inventory of what the solution collects and stores, and the requirement each item serves.
What this can look like
A list of the data fields the AML system holds and why each is needed.

5.11(a)(ii) Encryption and security controls

The AML solution shall:

implement security controls (including encryption of data at rest, in use and in transit, as appropriate) to protect the confidentiality, integrity, and availability of sensitive data.

What an examiner will look for
Security documentation and test reports showing encryption at rest and in transit and other controls.
What this can look like
Encrypted storage and connections, with security testing reports.

5.11(a)(iii) Role-based access control

The AML solution shall:

enforce role-based access controls, ensuring users access only the data and functions required for their duties.

What an examiner will look for
The roles defined in the system, what each can see and do, and the users assigned to them.
What this can look like
Analysts, reviewers, administrators and auditors have different roles with different rights.

5.11(a)(iv) Multi-factor authentication

The AML solution shall:

have secure authentication mechanisms, including Multi Factor Authentication (MFA), where appropriate.

What an examiner will look for
Authentication settings showing MFA enforced for users of the solution.
What this can look like
Every user signs in with a second factor.

5.11(a)(v) NDPA and data sovereignty

The AML solution shall:

support compliance with the Nigeria Data Protection Act (NDPA) and other relevant data protection and data sovereignty regulations and requirements.

What an examiner will look for
The data protection assessment for the solution, where data is stored and processed, and how NDPA obligations are met.
What this can look like
A data protection impact assessment covering the AML system and a record of where its data is hosted.

5.11(a)(vi) Recovery time and recovery point objectives

The AML solution shall:

Have defined Recovery Time Objective (RTO) and Recovery Point Objective (RPO) for AML systems, determined through a comprehensive Business Impact Analysis, and commensurate with the institution’s risk profile, size, complexity, and criticality of operations.

What an examiner will look for
The business impact analysis, the RTO and RPO set for the AML system, and recovery test results against them.
What this can look like
A business impact analysis sets how long the AML system may be down and how much data may be lost, and recovery tests check both.

5.11(b)(i) Data sovereignty and privacy compliance

The institution shall:

ensure that processing and storage of data by the AML Solution comply with Nigerian data sovereignty and privacy laws.

What an examiner will look for
The institution’s assessment of where the solution processes and stores data, against Nigerian data sovereignty and privacy laws.
What this can look like
A written assessment of data location and transfers for the AML system, approved by the data protection officer.

5.11(b)(ii) Retention and destruction

The institution shall:

implement appropriate data retention and destruction policies for AML/CFT/CPF data.

What an examiner will look for
The retention and destruction policy for AML data and evidence it is applied.
What this can look like
A policy that sets how long AML records are kept and how they are deleted after, with deletion logs.

5.11(b)(iii) Information security risk of the AML system

The institution shall:

assess and manage information security risks arising from the AML Solution as part of the institution’s broader cyber and information security framework.

What an examiner will look for
The AML solution in the institution’s information security risk register and assessments.
What this can look like
The AML system appears in the cyber risk register, with its own risks, controls and owners.

5.12 User Interface & Customisation

6 requirements

5.12(a)(i) Dashboards for metrics, alerts and cases

The AML solution shall:

provide dashboards and interfaces that support real-time or near real-time visibility of key AML/CFT/CPF metrics, alerts and case management workflows.

What an examiner will look for
A demonstration of dashboards showing current alert, case and metric figures.
What this can look like
A dashboard of open alerts, case ageing and key metrics, updated as work happens.

5.12(a)(ii) Usable interface for investigators

The AML solution shall:

offer a user-friendly interface that facilitates efficient navigation, analysis and decision-making by compliance, monitoring and investigation teams.

What an examiner will look for
A walk-through of the interface by the people who use it.
What this can look like
Analysts reach the customer, transactions and decision from the alert in a few steps.

5.12(a)(iii) Multi-entity, multi-currency, multi-jurisdiction

The AML solution shall:

where relevant, support multi-entity, multi-currency and multi-jurisdiction configurations, in line with the institution’s group structure and footprint.

What an examiner will look for
Where relevant, configuration by entity, currency and country that matches the group’s structure.
What this can look like
Rules and thresholds set per country and currency, and data kept separate per legal entity.

5.12(a)(iv) Configuration under change control

The AML solution shall:

support configuration of workflows, escalation paths, alert filters and other parameters under appropriate governance and change-control.

What an examiner will look for
The configuration settings for workflows, escalation and filters, and the change history and approvals for them. The Guidance Note names “change control” as part of governance (section 4(b)).
What this can look like
Workflow and filter changes are made in the system by authorised users and each change is recorded and approved.

5.12(b)(i) Configured to the institution

The institution shall:

deploy AML Solutions that are appropriately configured to reflect their business models, product portfolios, customer segments and risk profile, and are compliant with all applicable laws and regulations.

What an examiner will look for
The CBN assesses “the institution’s ability to demonstrate that its AML solution is appropriately configured to its risk profile” (Guidance Note, section 9), and compliance “including governance, configuration, and effectiveness” (section 3).
What this can look like
The configuration matches the products and segments the institution serves, with settings that differ where the risk differs.

5.12(b)(ii) Process for updating rules and settings

The institution shall:

maintain documented processes for reviewing and updating rules, scenarios, thresholds and other configuration settings in response to new risks, products, channels and typologies.

What an examiner will look for
The documented process and records of reviews and updates it produced. The Guidance Note names “change control” as part of governance (section 4(b)); change management and approval processes are a minimum governance element (Template, section 6.2).
What this can look like
A written process for when rules are reviewed (new product, new typology, scheduled review), who changes them and who approves.

6 Other Requirements

5 requirements

6(a) Vendor and third-party management policy

The institution shall:

maintain a Vendor/Third-Party Management Policy that covers procurement, implementation, support, change management, incident handling and exit strategies for AML Solutions.

What an examiner will look for
The policy, covering each of the listed stages. The roadmap asks for each vendor’s scope of work, dependency level and risk if delayed (Template, section 10.2).
What this can look like
A vendor policy applied to the AML solution provider, including an exit plan.

6(b) Third parties held to the standards

The institution shall:

ensure that third-party providers, including shared service centres and cloud providers, comply with all applicable provisions of these Baseline Standards and relevant CBN regulations.

What an examiner will look for
Contracts and assessments showing third parties are held to the relevant provisions. The Guidance Note says outsourcing “does not transfer” regulatory responsibility or “accountability for demonstrating compliance” (section 6).
What this can look like
Contract clauses and periodic assessments of the AML provider and its hosting provider against the relevant provisions.

6(c) Cybersecurity, shared services and operational risk rules

The institution shall:

comply with the CBN Risk-Based Cybersecurity Framework, Guidelines for Shared Services and any other applicable regulations relating to operational risks, as it pertains to IT risks, outsourcing and operational resilience.

What an examiner will look for
Evidence that the AML solution is covered by the institution’s compliance with the CBN Risk-Based Cybersecurity Framework and the shared services and outsourcing rules.
What this can look like
The AML system is in scope of the institution’s cybersecurity framework assessments and outsourcing register.

6(d) AI governance standard

The institution shall:

Adhere to ISO 42001 and/or any other such standards in the usage/governance of Artificial Intelligence as applicable to these Baseline Standards.

What an examiner will look for
Where AI is used: the AI governance standard adopted and evidence of following it.
What this can look like
An AI management framework aligned with ISO/IEC 42001 or a comparable standard, covering the AI used in AML.

6(e) Heightened due diligence on AML vendors

The institution shall:

Pending the issuance of dedicated third-party risk management regulations by the CBN, institutions shall apply heightened due diligence to vendors and third-party providers of AML Solutions, including assessment of the provider’s ability to support institution-specific configuration, provide access to audit and validation data, and meet the governance requirements of these Baseline Standards. Upon issuance of CBN third-party risk management regulations, those regulations shall apply in addition to and, where more stringent, in substitution of this provision.

What an examiner will look for
The due diligence file on the AML vendor, covering configuration, access to audit and validation data, and governance. The roadmap warns that responses should “not rely solely on vendor representations” (Template, section 2), and the Guidance Note says vendor claims of “full compliance” are not recognised (section 3).
What this can look like
A vendor assessment that tests configuration, audit-data access and validation support directly, rather than relying on the vendor’s claims.

Sources

  1. Central Bank of Nigeria, circular BSD/DIR/PUB/LAB/019/002, “Issuance of Baseline Standards for Automated Anti-Money Laundering (AML) Solution for Financial Institutions in Nigeria”, 10 March 2026, with the attached “Baseline Standards for Automated Anti-Money Laundering (AML) Solutions for Financial Institutions in Nigeria” (March 2026, Compliance Department).
  2. Central Bank of Nigeria, Compliance Department, letter CMD/DIR/PUB/CIR/001006, “Implementation of the Baseline Standards for Automated AML/CFT/CPF Solutions”, 31 March 2026, with the attached “Guidance Note on the Implementation of the Baseline Standards for Automated AML/CFT/CPF Solutions” and “CBN AML Baseline Standards Implementation Roadmap Template”.

Ask us

A question about this regulation?

A person at Archer reads every question and answers by email.

The Fraud & AML Library is a practical reference, not legal advice. What binds is the regulator’s own text, which every rule and regulation page links.

Get new rules and regulation changes by email

When we add a rule or a regulation changes, we’ll email you.

You can unsubscribe at any time. Privacy policy