Manage · risk-based approach

Approve faster where you safely can.
Defend every score.

Retire the spreadsheet. Score every customer against the four FATF dimensions with 240+ configurable rules, tier it to your own risk appetite, and trace any score back to the exact factors that produced it, so a low-risk customer moves fast and a high-risk one is caught, not averaged away.

30 minutes. Your data. No slides.
Risk assessment in Compliance Hub: the four FATF dimensions scored, with history and override
Score
Low risk · cleared to move fast
every factor traced · exported in one click
A global auction house A diamond house A watch maison A private bank A fund administrator names on request
What it is

A configurable, orchestrated FATF four-dimension scoring engine inside Compliance Hub that replaces the manual AML workbook one person maintains. You weight it, tier it, override it, and backtest every change before it goes live, and every score and change is an immutable audit event.

Why they chose it
"Onboarding, risk assessment and screening run with role-based permissions on one shared platform, so the tension between our AML obligations and the client experience our brand demands is finally resolved."
Managing Director and MLRO · global luxury diamond house · 8 years
Named reference available privately, on request, under client approval.

What you can do today

Shipped, non-AI, and running on your own data. Set your appetite, protect against the single high-risk factor, test a change before it moves a real customer, and show your working in one click.

Studio: dimension weights and the risk model
Your risk appetite, as configuration

Set your own risk appetite, and change it the same day

Score every customer across Customer, Geographic, Product or Service, and Delivery Channel, weight each dimension yourself, and edit the model in Studio as configuration, not a change request. 240+ configurable rules, and same-day changes rather than a multi-week vendor cycle.

See Studio
Score breakdown: highest dimension sets the tier
Never averaged away

Never let one high-risk factor slip through

A weighted composite produces the score, and the overall tier follows FATF guidance: the highest individual dimension sets the customer tier, so a single high-risk factor is never averaged away. The Geographic dimension starts from a 209-country database carrying 16 risk indicators each, including FATF grey and black-list markers.

See the FATF dimensions
Challenger backtest: the re-tiering delta before promotion
Test before it moves a customer

Change the model without risking your risk appetite

Run a challenger in shadow against live cases and backtest a proposed change across your whole historical book, so you see the re-tiering delta before a single real customer moves, then promote only when the evidence is in. Every change and promotion is one of 32 immutable audit event types.

See challenger testing
Evidence chain: factors, override reason, named user, timestamp
Defensible on demand

Show your working in one click

Every score carries an evidence chain back to the factors that produced it, and every override is recorded with a reason, a named user, and a timestamp, exported for a regulator from one record. Governed by role-based access, with four-eyes on high-risk changes.

See the audit trail

Under every score, an evidence chain you export for a regulator in one click. Through every change, a backtest and four-eyes sign-off. That is what having your back looks like.

4
FATF dimensions on every customer
240+
configurable rules you weight yourself
209
countries, 16 risk indicators each
32
immutable audit event types
Zenoo Labs · alpha and beta

A copilot that drafts the change, and pre-runs the backtest.

In Zenoo Labs we are testing a Risk Model Copilot, one of ten specialised AI agents, that suggests scoring changes and pre-runs the what-if backtest for you, plus a Screening Resolver that triages matches so the Customer dimension carries fewer false positives. In alpha and beta with design partners, human-in-the-loop and opt-in. Nothing goes live until a person approves the promotion.

AI proposes, a person promotes.
Labs: Risk Model Copilot and the what-if backtest
The result

A low-risk customer moves fast, a high-risk one is caught, and every score holds up when a regulator asks.

Same day

Change your model and promote it the same day, versus the multi-week cycle a vendor change request takes.

Questions

Risk scoring, honestly.

Does the risk model replace our existing AML workbook?

Yes. The Compliance Hub scoring engine is designed to replace the manual spreadsheet one person maintains. You configure the four FATF dimensions and their weights in Studio, run a backtest against your historical book before going live, and the workbook becomes an immutable audit trail rather than a living document somebody edits.

Can our compliance team change the risk model without involving Zenoo?

Yes. Studio's Risk Model Editor gives compliance teams direct ownership of the 240+ rules across the four FATF dimensions. You can test a proposed change against live cases with champion/challenger testing, run a what-if backtest across your historical book, and deploy with version history and rollback. No vendor change request is needed for routine rule updates.

What happens when a single high-risk factor appears in an otherwise low-risk customer?

The platform follows FATF guidance: the overall customer tier is set by the highest individual dimension, so a single high-risk factor is never averaged away by the composite score. You see both the weighted composite and the individual dimension scores in the evidence chain.

How does the Geographic dimension work?

The Geographic dimension starts from a 209-country database, each country carrying 16 risk indicators: FATF grey list, FATF black list, Financial Secrecy Index, Transparency International CPI, and others. You weight the Geographic dimension against the other three FATF dimensions in Studio and can override the indicator weights within it.

Can we export a full evidence chain for a regulator?

Yes. Every score carries a per-event audit record with the factors that produced it, any override reason, the named user, and a millisecond timestamp. You export that chain for a regulator from a single record rather than reconstructing it manually from portal logs.

What is the Zenoo Labs Risk Model Copilot?

The Risk Model Copilot is one of ten specialised AI agents in Zenoo Labs, currently in alpha and beta with design partners. It suggests scoring changes and pre-runs the what-if backtest for you. Nothing goes live until a person approves the promotion. It is opt-in, human-in-the-loop, and does not change your model autonomously.

Approve faster. Defend every score.

If your model does not hold up on your data, do not buy it. Bring your Monday inbox and we run it live in 30 minutes.

30 minutes. Your data. No slides.