At FNOL submission, extract the key fraud indicator fields: claimant identity, loss date, loss type, reported circumstances, involved parties, and policy information
Submit the claim data to a fraud analytics API (e.g., Verisk Insurance Services, LexisNexis Attract, or Shift Technology) for real-time fraud score generation
Parse the fraud score response: overall score, contributing factor flags (staged accident indicators, prior fraud flags, address inconsistencies, social network links), and recommended action
Route high-score claims to a Special Investigations Unit (SIU) queue with the fraud indicators pre-populated for investigator review
For borderline scores, apply secondary validation rules (e.g., ISO ClaimSearch prior claims check) before making a routing decision
Log all fraud scores and routing decisions with the claim record and track SIU referral outcomes to calibrate score thresholds over time
Known gotchas
Fraud scores used in claims handling may be subject to adverse action notice requirements under state insurance regulations; confirm whether the score constitutes a consumer report under applicable law
False positives in fraud scoring can result in delayed payments to legitimate claimants and expose the insurer to bad faith claims; set score thresholds conservatively and require human review before payment denial
ISO ClaimSearch and similar databases have reporting requirements; insurers that use the database are typically obligated to contribute their own claims data as a condition of access
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp