A life rating is a sum of named debits on one screen
Automated underwriting arrived with the treaty, but the schedule behind it is not yours to read or change. Decisimo puts that schedule in a scorecard and decision table your underwriting team can edit.
The rating is a sum of named debits in a scorecard
Each predictor has named bands, and each band carries its debits. The scorecard adds the debits and writes the total where the terms table reads it. Every predictor opens with a band for an unanswered application value that carries no debits, so the treatment is a visible row of the schedule. Underwriters change a band’s range, debits and order in the portal with no code, and every band, debit and term shown belongs to a demo case.
Rows turn the debit total into underwriting terms
Total debits and an open investigation determine the decision, extra mortality and review period together. A postponement is a row in the policy, so it is written with the other terms. Each rating row also reads the investigation column, so exactly one row matches a life. The percentage is extra mortality, never a premium.
A changed band becomes a release to test and read back
Every save creates a revision, and two revisions can be compared. The testing studio runs the decision flow on chosen values and shows each predictor’s value against its band, with debits and weight. A changed band becomes a new release on the same endpoint, which can require sign-off before deployment. Each decision’s input and output can be written to the insurer’s own Postgres database.
The obligation stays with the insurer
The EU AI Act treats risk assessment and pricing in life and health insurance as an Annex III high-risk use, with obligations from 2 December 2027. Life and health is the only insurance line Annex III names, and GDPR Article 22 applies to an automated decision today. See the governance obligations.
Questions we get asked
Can underwriters change a band and run tests or a backtest before deployment?
Underwriters can change bands in the portal with no code. The team can rerun saved test cases or run a backtest over archived decisions. Sign-off can be required per endpoint when a release is deployed.
How can the team read a rating back debit by debit?
The testing studio shows each predictor’s value against its band, with debits and weight. A live response includes that breakdown when tracing is requested; by default it carries the total and terms.
What happens when an application waits for a medical report?
The same scorecard and decision table can run inside a case. When a value it read is replaced, the decision task runs again; a human review comes back as a new task with current values, and the earlier review stays on the file. Learn about case work.
What happens when an application leaves a question unanswered?
The missing value falls into the band the underwriter gave it. Each predictor opens with a band carrying no debits for an unanswered question, so that treatment is visible in the schedule.
Can claims run on the same product?
Claims can run as cases, and standard risk can run in one call across the insurance industry. See insurance use cases.
Run your own debit schedule
Define your debit schedule in a scorecard and terms table, then read each revision in the portal.