ai Shipped · Flagship
AI Consistency Analyzer
Detects inconsistent decisions vs. similar prior claims, flags analyst drift, and checks policy-rule conformance on every decision.
Key benefits
- · Flags claims whose decision pattern diverges from similar prior cases
- · Surfaces analyst / reviewer-to-reviewer drift for QA
- · Policy-rule conformance check on each decision before it's finalized
- · Makes audit and appeals defensible
What it does
Every decision in a compensation program should be consistent — with the statute, with policy, and with similar prior decisions. VCPMS’s consistency analyzer runs three analyses automatically:
- Prior-case comparison. Compares the current claim against similar prior claims (similar crime type, similar benefit categories, similar collateral-source situation) and flags decisions that diverge substantially from the pattern.
- Reviewer drift. Tracks approval/denial patterns per analyst over time and surfaces drift — a signal for supervisor QA conversations, not auto-enforcement.
- Policy-rule conformance. Checks the decision against the tenant’s configured policy rules and flags non-conforming decisions before the Award is finalized.
What this means for programs
- Audit defensibility. When a reviewer’s decision is later questioned, the system can show the consistency analysis the reviewer saw.
- Training signal. New analysts who drift from team norms get surfaced in QA dashboards — before it becomes a pattern.
- Appeal preparation. Appeals staff have a ready view of comparable cases with similar decisions.
Constraint: advisory only
The analyzer flags; it does not block decisions. Overriding a flag requires a reason, which is captured on the record.