Scoring models
Scoring formalizes the assessment: questions are grouped, groups have weights, and certain answers block a positive result regardless of the score. The system fills in some answers itself from external sources and screening data.
Plan All plans.
What is inside
- Visual editor: categories, groups, questions, answer options and weights
- Two rule modes: a scored question and a blocking compliance criterion
- Stop factors and score ranges with a recommended risk level
- Answer auto-fill from an external source feature or screening
- Model versions: a check result is bound to its version for good
- Dry run of a new version against historical checks before publishing
- Model application conditions by subject type and basis
- Mandatory justification for answers that lead to a rejection
Boundaries
Only users with a separate permission edit models, and by default the specialists who run checks do not have it. A published version cannot be changed, only replaced by the next one.
Frequent questions
How is a stop factor different from a blocking criterion?
A stop factor fires on a specific answer inside scoring. A blocking criterion is checked before the full check: if it fails, the check does not start, and the company does not pay for external source requests.
How do we know a new model will not break decisions?
Before publishing, the version is run against completed checks. You see how many decisions would change and in which direction.
Can a model be used for declarations?
Yes. A published model version serves as the form of a conflict of interest declaration campaign.
Use cases with this feature
Other features
Shall we bring order to your checks?
We will show the dossier, a check, a verdict and monitoring on a demo environment, review your regulations and calculate a pricing plan for your actual volume.