VERDICTS

STOP, FIX or SCALE.

How the engine rules, project by project, based on use, sector and risk. The sector doesn't set the verdict; the deployment conditions do.

Illustrative examples
Verdict
Sector
Finance Stop

Credit scoring

Decision to makeLet AI decide loan approval on its own?
Main reasonA high-impact automated decision on individuals, without sufficient safeguards or human recourse.
Minimum conditionEffective human oversight and a defensible explanation for every rejection.
TrailAI Act · GDPR · ISO 42001
HR Stop

CV screening

Decision to makeAutomate candidate shortlisting?
Main reasonRisk of discrimination on sensitive data, with no bias controls in place.
Minimum conditionDocumented bias controls, with a human in the loop at every step.
TrailAI Act · GDPR · ISO 42001
Health Fix

Diagnostic support

Decision to makeUse it in real clinical settings?
Main reasonReal clinical value, but oversight and validation still incomplete.
Minimum conditionClinical validation and formal medical oversight before production.
TrailAI Act · GDPR · ISO 42001
Insurance Fix

ML pricing

Decision to makeDeploy automated pricing?
Main reasonSolid performance, but explainability and pricing fairness still need to be proven.
Minimum conditionExplainable pricing factors and documented fairness testing.
TrailAI Act · GDPR · ISO 42001
Industry Scale

Predictive maintenance

Decision to makeRoll out to every plant?
Main reasonGovernance, oversight and performance in place, limited human impact.
Minimum conditionPhased rollout with drift monitoring.
TrailAI Act · GDPR · ISO 42001
Agents Fix→Scale

AI sales agent

Decision to makeLet it act on the CRM and engage customers?
Main reasonUseful autonomy, but its action scope and human oversight still need clear limits.
Minimum conditionA bounded action scope, with human sign-off on sensitive actions.
TrailAI Act · GDPR · ISO 42001

No example matches this filter.

Does your project look like one of these? Submit a similar project →