Skip to main content
Back to challenges
Challenge

Build an Automated Insurance Claims Fraud Agent with FastHTML and Instructor

Insurance claims processing in Asia-Pacific markets requires fast automated fraud triage to combat inflated or fraudulent loss reports. Construct a web dashboard and structured extraction agent using FastHTML and Instructor in Python that screens claim documentation.

CybersecurityHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Build an interactive FastHTML web portal connected to an Instructor structured LLM agent for insurance claim fraud detection.

Delivery guide

How work is evaluated

Evaluation

Evaluates accuracy of claim fraud classification and Pydantic schema adherence.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

test_siu_referral

Ensure policy inception proximity triggers SIU referral

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

f1_score

F1 score for claim fraud triage • target: 0.95 • range: 0.85-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Build a pure Python FastHTML UI for submitting and viewing insurance claims.

  • Use Instructor to enforce structured Pydantic schemas on claim narrative extractions.

  • Calculate fraud likelihood based on repair cost outliers and treatment timestamps.

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic insurance claim records and accident incident descriptions.

Challenge setup

Evaluates accuracy of claim fraud classification and Pydantic schema adherence.

Challenge input

JSON claim submission including description, claimed_amount, and policy_age_days

FastHTML

UI framework requirement

Instructor

Structured extraction tool requirement

Evaluated output

JSON validated model with fraud_score, routing_decision, red_flags

Evaluation checks
  • Ensure policy inception proximity triggers SIU referral
  • F1 score for claim fraud triage • target: 0.95 • range: 0.85-1
Configured evaluation evidence
  • Benchmark: Insurance Triage Benchmark v1
  • F1 Score target: 0.95
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Action Space
  • Instructor · Required

Evaluation contract

  • test_siu_referral · Weight 1
  • f1_score · Weight 1

Metrics

  • f1_score · Target 0.95

Benchmarks

  • Insurance Triage Benchmark v1
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON claim submission including description, claimed_amount, and policy_age_days

    Output: JSON validated model with fraud_score, routing_decision, red_flags

    Execution harness: Python

Evaluation tools

  • FastHTML · web_framework
  • Instructor · structured_llm_tool

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build an Automated Insurance Claims Fraud Agent with FastHTML and Instructor