Clifford Bernard

CaseC-0002

Burst pipe, flooded basement

A clean home claim sent to a specialist until the model's answers were weighed properly.

home claim · $18,400

What the claimant wrote

Home Water Damage, basement flooding from burst pipe. Supply line failed suddenly; plumber report confirms failed fitting. Flooring, drywall and furniture in the finished basement ruined. Restoration estimate $18,400.

At a glanceEvery version

This claim through all ten versions

0Today

By hand

The claim joins one queue with every other claim and waits about 5.0 days for an adjuster to read it, check the policy and decide.

W0 to W3: turning the answers into a decision

  1. W0 · Trust the one big answer

    Sent to a person: flagged as needing a specialist. The right call

    The steps it went through
    1. Code R-HV-50K not fired: Claims over $50,000 require mandatory human review
    2. Code R-INJECTION not fired: Suspected manipulation of automated processing quarantines the claim
    3. Code R-FRAUD not fired: Claims with fraud indicators require investigation
    4. Code R-POLICY not fired: Claims on a policy that is not in force go to a human (denials are never automatic)
    5. Code R-TYPE-CONFIDENCE not fired: Uncertain claim type cannot be routed automatically
    6. Code R-SPECIALIST FIRED: Claims needing domain expertise are assigned to a specialist Decided here
    7. Person Disposition human_review by R-SPECIALIST

  2. W1 · Combine the small answers

    Sent to a person: flagged as needing a specialist. The right call

    The steps it went through
    1. Code Claim C-0002: $18,400, policy active.
    2. Decision model The model answers fourteen typed questions about completeness, evidence and fraud; code decides, it never does.
    3. Code Claims over $50,000 require mandatory human review
    4. Code Suspected manipulation of automated processing quarantines the claim
    5. Code Claims with fraud indicators require investigation
    6. Code Claims on a policy that is not in force go to a human (denials are never automatic)
    7. Code Claims needing domain expertise are assigned to a specialist Decided here
    8. Person Sent to a person: R-SPECIALIST.

  3. W2 · Learn how much to trust each answer

    Paid automatically. The right call

    The steps it went through
    1. Code Claim C-0002: $18,400, policy active.
    2. Decision model The model answers fourteen typed questions about completeness, evidence and fraud; code decides, it never does.
    3. Code Claims over $50,000 require mandatory human review
    4. Code Suspected manipulation of automated processing quarantines the claim
    5. Code Claims with fraud indicators require investigation
    6. Code Claims on a policy that is not in force go to a human (denials are never automatic)
    7. Code Fourteen learned weights score approval at 98.3% and a request for information at 0.5%; the higher one is proposed.
    8. Code Certain enough, so it is handled with no person. Decided here
    9. Code Handled with no person: paid.

  4. W3 · Be stricter as amounts rise

    Paid automatically. The right call

    The steps it went through
    1. Code Claim C-0002: $18,400, policy active.
    2. Decision model The model answers fourteen typed questions about completeness, evidence and fraud; code decides, it never does.
    3. Code Claims over $50,000 require mandatory human review
    4. Code Suspected manipulation of automated processing quarantines the claim
    5. Code Claims with fraud indicators require investigation
    6. Code Claims on a policy that is not in force go to a human (denials are never automatic)
    7. Code Fourteen learned weights score approval at 98.3% and a request for information at 0.5%; the higher one is proposed.
    8. Code The certainty needed grows with the amount at stake: doubt is scaled by the claim's value against a $5,000 reference, giving 93.8% confidence.
    9. Code Certain enough, so it is handled with no person. Decided here
    10. Code Handled with no person: paid.

Around W3: the work people still do, and what else was tried

  1. + A quick confirmation for near-certain claims (now for requests only)

    Paid automatically. The right call

    The steps it went through
    1. Code The fitted workflow handles this claim with no person; no lane is needed.
    2. Code Handled automatically; this claim never reaches a lane.

  2. + Grade the fraud response

    Paid automatically. The right call

    The steps it went through
    1. Code The fitted workflow handles this claim with no person; no lane is needed.
    2. Code Handled automatically; this claim never reaches a lane.

  3. + A second vote on near-misses (a negative result)

    Paid automatically. The right call

    The steps it went through
    1. Code The fitted workflow handles this claim with no person; no lane is needed.
    2. Code Handled automatically; this claim never reaches a lane.

  4. + Ask the customer one question

    Paid automatically. The right call

    The steps it went through
    1. Code The fitted workflow handles this claim with no person; no lane is needed.
    2. Code Handled automatically; this claim never reaches a lane.

  5. + Watch after the decision

    Paid automatically. The right call

    The steps it went through
    1. Code The fitted workflow handles this claim with no person; no lane is needed.
    2. Code Handled automatically; this claim never reaches a lane.