Claims / SIU
A referral, never a determination
Fraud rules and predictive screening raise candidates. An investigator works them. A claim cannot close while the referral is open.
The platform ships a rule catalog with nothing enabled, because a rule tuned on somebody else’s book produces referrals your team learns to ignore.
Review the rule catalogDoes the system decide a claim is fraudulent?
No, and there is no configuration that would let it. Screening scores claims against your rules and raises a referral for a human investigator. A determination is an outcome of that investigation — conducted by a person, recorded on the file, with their name on it.
Your rules, your thresholds
From signal to outcome
- 01
Rules are yours, and inert by default
The catalog ships with nothing switched on. You enable what matches your experience, at thresholds you can justify if asked why.
- 02
Screening raises candidates
Rule scores and, where entitled, predictive screening produce referral candidates — with the reasoning recorded on the claim.
- 03
Referral is an overlay
The SIU role sits on top of an adjuster’s role rather than replacing it, so investigation is not a separate system to reconcile afterwards.
- 04
Closure is blocked
A claim cannot close while an SIU referral is open. The transition refuses rather than warning, which is the whole point.
- 05
Outcome is recorded
Investigation and result live on the claim, so a pattern of referral and a pattern of outcome are both reportable.
Why defaults ship off
Two failures that arrive together
A mis-tuned fraud rule does not just waste time. It creates a documented pattern of accusation against a class of claim, which is a question you may have to answer.
| Consequence | Why it matters |
|---|---|
| Referral fatigue | A team that learns to dismiss referrals misses the real ones |
| Indefensible patterns | Systematic referral of a claim type is a question you may be asked to justify |
| Enabled defaults | Not shipped — the catalog is inert until you configure it |
| AI screening | Available where entitled; produces a referral, never a conclusion |
| Sanctions screening | Separate: OFAC results and overrides recorded against the party |
What the platform will not conclude
It routes and records. Every judgment in this area belongs to a person, deliberately.
No fraud determination
Not by rule, not by model, not by configuration. A referral is the ceiling of what automation produces here.
Thresholds are yours to defend
We enforce the thresholds you set. Whether they are reasonable is a question for you, and the record will show what they were.
Investigation capacity is human
Better screening raises more candidates. Without investigators, that is a longer queue rather than a better outcome.