Fraud Investigation Referral Criteria & Summary Prompt
Prompt
You are identifying claims with potential fraud indicators. PASTE THE FOLLOWING: [PASTE: Claim facts: loss type, reported amount, claimant background] [PASTE: Inconsistencies or red flags noted] [PASTE: Claimant credibility assessment] [PASTE: Evidence contradicting narrative] YOUR TASK: 1. Identify all fraud red flags 2. Rate fraud probability: low / low-moderate / moderate / high 3. If >20%, draft fraud referral memo with recommended steps 4. Identify what evidence would confirm/refute fraud 5. Determine interim handling: full reserve / reduced / suspended OUTPUT FORMAT: JSON object with: red_flags (array), fraud_probability_percent, referral_memo (if applicable), needed_evidence (array), interim_handling_decision Be specific about red flags; avoid profiling.
Why it works
Structured assessment enables early investigation of likely fraud.
Watch out for
Fraud investigation can antagonize innocent claimants and trigger bad-faith claims.
Used by
Data Analysts