Risk scoring, defined, and why it is not our product

Solidus has no risk engine, and a portable credential contributes nothing to risk scoring. This page defines the term because our buyers use it constantly, and then says where the boundary is.

What it is

Assigning a risk rating to a customer or a transaction, so that effort can be concentrated where it matters. A rating typically combines:

• Who the customer is, jurisdiction, occupation, whether they are a politically exposed person, ownership structure for an entity.
• What they do, transaction patterns, volumes, counterparties, how the behaviour compares to what was expected at onboarding.
• What is known externally, sanctions exposure, adverse media, prior filings.

The output drives everything downstream: how much diligence at onboarding, how often periodic review happens, what gets escalated to a manual review queue, and where a false-positive rate is worth tuning.

Why a credential does not help, stated rather than finessed

A verifiable credential attests that an identity was verified. It carries no information about risk. Knowing a person is who they say they are tells you nothing about whether their transaction pattern is unusual, whether they appear on a list, or whether their jurisdiction warrants enhanced diligence.

There is a tempting sentence available here, something about verified identity being a stronger input to a risk model, and we are not going to write it, because the honest version is that a risk model's identity input is rarely the weak link. The weak link is behaviour data and external exposure, and we have neither.

Risk scoring is somebody else's product category. If that is what you are evaluating, evaluate them.

The one thing worth saying

Where identity infrastructure does matter to a risk programme is upstream and narrow: a verification that can be re-presented rather than re-run reduces friction in the parts of your cycle that are triggered by risk, most concretely in re-KYC, where the repetition is mandatory. That is an argument about cost and customer friction in the workflow around scoring, not about the scoring itself, and it is the only claim we will make on this page.