Knowledge Of The Art Ltd

AI-Enabled AML/CFT/CPF — Risk Intelligence, Human Judgement and Accountable Compliance

In 2027 Mauritius will be assessed on whether its AML/CFT/CPF framework works, not on whether it exists. The assessment methodology has moved from technical compliance to demonstrated effectiveness, and supervisory posture has moved with it. The real challenge is no longer whether compliance functions will use AI. It is whether they can use it lawfully, at scale, and prove afterwards exactly where the machine stopped and the professional began.

This practical masterclass takes MLROs and Compliance Officers beyond tools and prompts. Participants will learn to extend detection across an entire client and structure population, to hold the human decision gate where a transaction is legally possible but regulatorily unacceptable, and to assemble the evidence that proves adequate procedures under section 52 of the Financial Crimes Commission Act. And, all in line with actual and known-forthcoming Mauritian Laws and International best practices.

Through live demonstrations, detection drills and compliance-focused working labs, participants will screen a fiduciary book beyond the United Nations lists, unwind layered ownership to a controlling interest, record decisions as they are taken, and defend the adequacy of their procedures under adversarial challenge.

Participants will learn how to:

• Build a dated forward regulatory register mapping AMLA 2026, the ESAAMLG assessment criteria and anticipated reform to their own institution's exposure.
• Determine where AI may assist, where human review is required, and where the decision must remain entirely human.
• Apply an AI Permission Map defining lawful tool use by data class under the Data Protection Act 2017 and FSC Guidance Notes No. 4 on the responsible use of AI.
• Deploy AI-assisted detection across a full population of administered entities rather than a sample, producing a reasoned triage list with grounds stated for every escalation.
• Reconstruct source of wealth and source of funds across layered chains, nominee arrangements and circular flows.
• Screen against consolidated sanctions and proliferation-financing designations beyond the United Nations lists, and act within the 24-hour freezing obligation.
• Distinguish a defensible alert from a false positive, and manage alert fatigue at population scale.
• Record decisions at the moment they are taken, with evidence typed STATED, DERIVED or FRAMED.
• Run the supervisory examination on their own file before a regulator does, against the five evidential principles of the FCC Guidelines on Legal Persons.
• Assemble and defend a section 52 adequate-procedures file under challenge.

The Detection and Defence Cockpit
A live financial-crime intelligence dashboard demonstrating:

• Population coverage and reasoned triage indicators
• Consolidated sanctions and proliferation-financing screening status
• Ownership-chain resolution and controlling-interest exposure
• Human overrides, escalation triggers and decision ownership
• Evidence typing across machine contribution and human conclusion
• Periodic-review trigger monitoring and ageing
• Board and committee-ready governance reporting

Participants will see how a fragmented client book can be converted into a structured oversight system that supports management action while preserving human accountability.

Live Lab 1 — The Detection Room

• Structure sweep across a portfolio of administered entities and a fund
• Consolidated sanctions and proliferation-financing screening drill
• Ownership-chain unwinding to a designated controlling interest, within the statutory freeze window
• Counterparty forensics on an operating investee, with the limits of the models made explicit

Live Lab 2 — The Defence File

• The decision gate: a legally possible but regulatorily unacceptable transaction
• Evidence typing at the moment of decision
• Self-inspection drill against the five FCC evidential principles
• Section 52 hearing — adequate procedures defended under adversarial challenge, with a prepared challenge-response pack

Target Audience: MLROs, Deputy MLROs, Compliance Officers, Data Protection Officers, resident Directors and senior fiduciary staff of FSC-licensed management companies, fund administrators, CIS managers, and corporate and trust service providers. Also relevant to internal auditors, company secretaries and senior managers with AML/CFT/CPF oversight responsibilities.

Participants should bring their fully charged laptop and have access to a paid version of their preferred AI platform. Claude is recommended for selected practical exercises. No prior experience of artificial intelligence is required. No live client data is processed at any point. All laboratory work is conducted on a fictional fiduciary case universe built for the purpose.

These are not passive demonstrations. Participants will interrogate evidence, challenge AI outputs, reject what the model gets wrong, make decisions and document the reasoning required to stand behind them.

Schedule: Date: 28 October 2026 | Time: 0815 – 1635 | Venue: Le Jadis Luxury Resort

Standards: Professional, practical, real-life, up-to-date action-packed seminar. Training centre and courses approved by the Mauritius Qualifications Authority and fees are refundable by HRDC for company-sponsored participants. 

Cost of training: Rs. 25,000 (MQA-Approved and HRDC Refundable up to 90%) - Net Investment after HRDC Refund: Rs.2,500.

CLICK HERE TO DOWNLOAD OUR BROCHURE

CLICK HERE TO APPLY ONLINE

FOR MORE INFORMATION, PLEASE LEAVE A MESSAGE ON OUR CONTACT US PAGE ON THIS WEBSITE OR PHONE US ON 466 3333

What is the AI-Enabled AML/CFT/CPF Masterclass?
It is a one-day, practical masterclass for MLROs, Compliance Officers and directors of Mauritian management companies and fund administrators. It teaches how to apply artificial intelligence lawfully across detection, sanctions screening and evidence assembly, while keeping the compliance decision in human hands and proving it afterwards.

When and where will it take place?
Wednesday 28 October 2026, from 0815 to 1635, at Le Jadis Luxury Resort, Mauritius.

Who should attend?
MLROs, Deputy MLROs, Compliance Officers, Data Protection Officers, resident directors and senior fiduciary staff of FSC-licensed management companies, fund administrators, CIS managers and corporate and trust service providers.

Is the course MQA-approved and HRDC-refundable?
Yes. The course is MQA-approved and qualifies for an HRDC refund of up to 90% for eligible company-sponsored participants, subject to applicable HRDC conditions.

What makes this different from other AML/CFT training in Mauritius?
Participants do not leave with notes. They leave with four completed instruments built during the day in their own institution's context: a forward regulatory register, an AI Permission Map, a decision record in typed-evidence format, and the skeleton of a section 52 defence file. It is also the first AML/CFT/CPF programme in Mauritius built around the lawful use of artificial intelligence on the client file.

Do I need prior experience with artificial intelligence?
No. Participants work with the AI tool their own institution already licenses. The programme is deliberately tool-agnostic and teaches method rather than proficiency in a named system.

What is a section 52 defence file?
Section 52 of the Financial Crimes Commission Act requires every legal person in Mauritius to maintain adequate procedures to prevent financial crime committed on its behalf, with a fine of up to MUR 20 million. The defence is evidential: the legal person must prove, on the balance of probabilities, that adequate procedures were in place. A defence file is the assembled evidence that would satisfy that burden.

Does the course cover proliferation financing?
Yes. A full module addresses countering proliferation financing, including targeted financial sanctions under the United Nations Sanctions Act 2019, FATF Recommendations 6 and 7, the 24-hour freezing obligation, dual-use goods indicators, and screening beyond United Nations designations.

Will client data be used during the training?
No. All laboratory work is conducted on a fictional fiduciary case universe. Participants apply the methods to their own book after the programme, within the AI Permission Map they build during it.

Is the course relevant to the 2027 ESAAMLG evaluation?
Yes. The programme is built around the effectiveness-based assessment methodology, which tests whether controls produce outcomes rather than whether frameworks exist on paper. 2027 is Mauritius's preparation year.

What will participants learn?
Participants will learn to build a forward regulatory register, apply an AI Permission Map, deploy AI-assisted detection across a full client population, screen for sanctions and proliferation-financing exposure within statutory deadlines, record decisions with typed evidence, and assemble and defend a section 52 adequate-procedures file.