MODULE · ADD-ON

Muhafız — the AML/MASAK investigation copilot

Muhafız is an add-on that accelerates AML alert triage and case-narrative drafting. The decision always rests with your investigator; Muhafız prepares context, it does not judge.

Assists with alert triage.
Drafts case narratives.
Every step is logged; human control is essential.

Use cases

Alert triage

Suspicious-activity alerts are contextualized with customer history, linked accounts, and transaction patterns — the investigator decides what to look at first on firmer ground.

Case-narrative drafting

A sourced draft of the case narrative underlying a suspicious-transaction report is prepared; the final text always comes from the investigator’s hand.

Customer profile summary

Linked accounts, transaction patterns, and past alerts are compiled on one summary screen with sources; case-opening time shortens.

Case closure

Justification documentation for a closed case is prepared from templates, with sources; the trail of the closure decision is preserved for audit.

Capabilities

Investigation support

  • Adds context to the alert; it does not decide or impose conclusions
  • Summarizes linked transactions and accounts with sources
  • Works on in-house AML data; sends nothing outside

Narrative drafting

  • Sourced, editable case-narrative drafts
  • Structure aligned with your reporting templates
  • Every draft requires human approval; there is no unapproved output

Trail keeping

  • Every AI contribution to an investigation is logged
  • Human decisions and AI suggestions are recorded separately
  • Traceable input to MASAK process documentation

How it works

  1. 01

    An alert arrives

    Alerts from your monitoring system connect into Muhafız.

  2. 02

    Context is compiled

    Customer history, linked accounts, and patterns are summarized with sources.

  3. 03

    The investigator decides

    Triage and go/no-go belong to the human; Muhafız offers context, not verdicts.

  4. 04

    A narrative draft is produced

    If the case proceeds, a sourced case-narrative draft is prepared.

  5. 05

    It is recorded with human approval

    The text the investigator approves is archived with the full trail.

What it delivers

  • Designed to support reducing the effort spent per case
  • Human control is fundamental: the decision is always the investigator’s
  • Every step is logged; the process can be explained in a MASAK audit
  • AI support for AML work without data leaving the institution

Why it matters

In AML, AI creates value only while preserving human control. Muhafız does not decide; it prepares the context and the draft that speed up the investigator’s own decision.

Auditability

Investigation steps, AI contributions, and human approvals are recorded separately and auditably.

Related modules