Risk rating

The engine turns form answers and analyst inputs into a risk rating on a 0 to 100 scale and a risk class: Low, Medium, or High. The number stays truthful; escalation acts on the classification.

How a rating is produced

Applicant answers

Analyst inputs and confirmations

Weighted risk criteria

Risk rating 0 to 100

Risk class from configurable bands

Mandatory escalation triggered?

Yes

Force class to High

Class from the bands

What makes the rating yours

Three things make the rating both tenant-specific and trustworthy.

Configurable bands

Each risk engine defines where Low, Medium, and High sit, so your own methodology drives the classification rather than a fixed scale.

Mandatory escalations

An engine can declare conditions that force the class to High regardless of the number, for example a counterparty in a sanctioned or FATF-listed jurisdiction.

Mandatory analyst sign-off

Criteria a form cannot answer are surfaced to the analyst, who confirms or overrides each one before approval is possible.

On escalation, the number stays truthful and only the classification escalates, so you keep an honest score and a conservative outcome.

Some criteria cannot be answered from a form: a relationship-tenure judgement, a screening outcome, a partner-risk assessment. These are surfaced to the analyst, who confirms or overrides each one. An application cannot be approved until every required analyst input is recorded.

Preliminary and confirmed ratings

Where an intake is analyst-scored, the applicant's own answers still produce a preliminary rating on submission, so your analyst starts from an informed position rather than a blank form. The analyst then reviews each criterion and confirms or overrides it.

A confirmed value always wins over the preliminary, and approval stays blocked until every criterion is confirmed.

Where analysts do this

The risk evaluation tab in the dashboard shows the rating, the per-category breakdown, the sign-off panel, and per-category override with a mandatory reason. See The compliance dashboard.

Fail-closed guarantees

Every path that could quietly produce a favourable answer produces a conservative one instead.

Situation	Behaviour
A screening call errors	Recorded as failed. Never read as clean.
A scoring input is missing	Fails closed at the criterion maximum. Never a silent zero.
The adverse-media provider is unavailable	That dimension is excluded and flagged “not assessed” rather than counted as clean.
A scoring rule cannot be resolved	The rating is marked as failed and the class is forced to High.

The rating is frozen at first evaluation, so a rating your analyst approved cannot silently shift when you later edit the engine. See Core concepts.