Product walkthrough

From alert to goAML STR, with a human approving every step.

An AI case agent for UAE-regulated VASPs and fintechs. This walkthrough follows one case through every screen of the product, from sign-in to the regulator's request.

Abu Dhabi · September 2026 All customer data shown is synthetic 25 screens
01Triage

Alerts grouped into cases, prioritised with an SLA.

Agent
02Investigate

KYC, transactions, screening and typologies in one case file.

Analyst + agent
03Draft

Narrative with citations, confidence and a recommended disposition.

Agent
04Approve

Analyst edits, MLRO approves. Four eyes on every filing.

Analyst + MLRO
05File

goAML STR built and schema-validated before submission.

MLRO
06Prove

Regulator RFIs answered from a replayable audit log.

MLRO + auditor
01

Get started

Sign in, set up the organisation, connect data sources.

01 · Get started

Sign-in and workspace

  1. Single sign-on (SSO / SAML 2.0); no separate passwords to manage.
  2. One workspace per reporting entity: licence, regulator and FIU profile set once.
  3. Data stays in the UAE region and is never used to train models.
Sign in
Sign-in screen with single sign-on

01 · Get started

Connecting data sources

  1. Connect the tools already raising alerts: Sumsub, ComplyAdvantage, Chainalysis, Elliptic, TRM, Fireblocks.
  2. Core ledger via API or CSV upload for fiat flows.
  3. Every connected source becomes a citable reference in later narratives.
Settings · Data sources
Data sources screen listing connected screening and custody providers
02

Triage

Alerts arrive from screening; FinAML groups them into cases and sets priority.

02 · Triage

Alert queue

  1. Alerts from every source land in one queue, sorted by risk and SLA.
  2. Related alerts on the same customer are grouped into a single case.
  3. Analyst assigns or opens a case; the suggested priority can be overridden.
Alerts
Alert queue sorted by risk and SLA
03

Investigate

One case file: customer, transactions, screening hits and matched typologies.

03 · Investigate

Case workspace

  1. Left rail: key facts, typology matches with confidence, investigation checklist.
  2. Centre: grouped alerts, funds-flow graph and counterparties with screening results.
  3. Right: the case agent, with narrative draft, cited sources and recommended disposition.
Cases · CASE-2026-00417
Case workspace with alerts, funds-flow graph and the case agent panel

03 · Investigate

Customer 360

  1. Profile: KYC tier, licence, onboarding documents in Arabic or English.
  2. Beneficial owners with ownership chain and PEP / sanctions screening.
  3. Risk history: rating changes, prior cases and their outcomes.
Customer · Profile
Customer 360 profile
Customer · Beneficial owners
Beneficial owners with ownership chain and screening results

03 · Investigate

Transactions: table and flow graph

  1. Table view: on-chain and fiat transfers with counterparty risk and flags.
  2. Flow graph: the same data as a funds-movement pattern, exposure highlighted.
  3. Selected transactions carry straight into the STR transaction list.
Transactions · Table
Transactions table with counterparty risk
Transactions · Flow graph
Funds-flow graph with exposure highlighted

03 · Investigate

Typology library

  1. Library of UAE-specific typologies, each mapped to VARA, FSRA, DFSA or CBUAE guidance.
  2. Detail page: indicators, thresholds and the cases where it matched.
  3. Custom typologies can be added for the entity's own risk assessment.
Typologies
Typology library
Typologies · T-07 Layering via mixer exposure
Typology detail page with indicators and thresholds
04

Draft and decide

The agent writes the narrative with citations; the analyst edits; the MLRO decides.

04 · Draft and decide AI output

Narrative draft

  1. Every sentence carries a numbered citation to the alert, transaction or document behind it.
  2. Confidence level and the sources used are shown next to the draft.
  3. The analyst edits inline and picks a disposition; edits are stored as a diff.
Case · Narrative and disposition (analyst)
Narrative draft with citations and recommended disposition

04 · Draft and decide

MLRO approval

  1. Four-eyes: the MLRO reviews the analyst's decision and the edits made to the draft.
  2. Approve, request changes with a comment, or reject.
  3. Each decision writes an immutable, hash-chained audit entry.
Approvals · CASE-2026-00417 (MLRO)
MLRO approval view showing the analyst's edits as a diff
05

File

A goAML STR, schema-validated and approved before anything is submitted.

05 · File

STR builder (goAML)

  1. Report, subjects, transactions and narrative are pre-filled from the case.
  2. Each transaction is mapped to the goAML fields; gaps are flagged before submission.
  3. Validation runs against the current UAE FIU schema.
STR builder · Transactions
goAML STR builder, transactions step

05 · File

goAML XML preview

  1. The exact XML that will be submitted, with schema validation results.
  2. Submit stays disabled until validation passes and the MLRO has approved.
  3. Download for manual goAML upload, or submit directly where the integration is enabled.
STR builder · XML preview
goAML XML preview with validation results

05 · File

Approvals on mobile

  1. Approvals queue: what is waiting on the MLRO, with SLA countdown.
  2. Approve STR: summary, sources and the analyst's edits. Approve or request changes.
  3. Confirmation with the filing reference and audit entry hash.
Approvals queue
Mobile approvals queue
Approve STR
Mobile approve STR screen
Confirmation
Mobile confirmation with filing reference
06

Respond and prove

Regulator requests, a replayable audit log, and the controls behind the agent.

06 · Respond and prove

Regulator requests

  1. RFIs from VARA, ADGM FSRA, CBUAE or the FIU tracked with deadlines and owners.
  2. Response pack assembled from linked cases, STRs and documents.
  3. Same four-eyes approval before anything leaves the organisation.
Regulator requests
Regulator requests list with deadlines
Regulator requests · Detail
Regulator request detail with response pack

06 · Respond and prove

Audit log and decision records

  1. Append-only, hash-chained log of every human and model action.
  2. Decision record: prompt version, sources used, raw model output, human edit as a diff, approver.
  3. Read-only auditor role; records anchored on ADI Chain for tamper evidence.
Audit log
Audit log list
Audit log · Decision record
Decision record with prompt version, sources and human edit diff

06 · Respond and prove

MLRO analytics

  1. Case volume, time per case and SLA performance by analyst and typology.
  2. STR outcomes and false-positive rate over time.
  3. Agent acceptance rate: how often recommendations are accepted, edited or overruled.
Analytics · MLRO overview
MLRO analytics overview

06 · Respond and prove

Team, roles and agent policy

  1. Roles: Analyst, MLRO, Auditor. Separation of duties enforced, not just documented.
  2. Agent policy: what the agent may do unattended and what always needs a person.
  3. Model and prompt versions pinned; changes are logged like any other decision.
Settings · Team and roles
Team and roles settings
Settings · Agent policy
Agent policy settings

Second-line AML, done in hours instead of days.

Pilots for VARA-, ADGM- and CBUAE-licensed entities are open now.

[email protected]