Build a Cooperative Audit and Governance Engine with CrewAI
Develop a multi-agent cooperative auditing system tailored for ASEAN student and regional co-ops. Using CrewAI, assign specialized role-playing agents (Auditor, Bookkeeper, Compliance Officer) to review financial transaction ledgers and generate regulatory reports automatically.
What you are building
Orchestrate role-playing CrewAI agents to perform end-to-end audit compliance and fraud detection for student cooperative accounts.
How work is evaluated
Evaluate accuracy of agent anomaly detection in cooperative transaction ledgers.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
flagged_anomaly_correctness
All non-compliant transactions accurately flagged
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
audit_accuracy
Precision score of audit classification • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Define multi-agent roles, goals, and backstories using CrewAI Agent class
Build sequential and hierarchical crew task workflows for ledger parsing
Implement automated anomaly detection tools for expense reimbursement claims
Generate regulatory compliance scorecards formatted for cooperative registrar standards
Resources
Simulated cooperative financial ledger dataset featuring standard and anomalous entries.
Challenge setup
Evaluate accuracy of agent anomaly detection in cooperative transaction ledgers.
Challenge input
JSON containing list of transactions with fields id, amount, category, user
CrewAI
Framework for multi-agent role-playing workflows.
Evaluated output
JSON containing flagged_transactions, compliance_passed, summary
- All non-compliant transactions accurately flagged
- Precision score of audit classification • target: 0.95 • range: 0-1
- Benchmark: FinAudit_Crew_Bench
- Audit Accuracy target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- CrewAI · Required
- crewAI · Optional
- Zed · Optional
Evaluation contract
- flagged_anomaly_correctness · Weight 1
- audit_accuracy · Weight 1
Metrics
- audit_accuracy · Target 0.95
Benchmarks
- FinAudit_Crew_Bench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing list of transactions with fields id, amount, category, user
Output: JSON containing flagged_transactions, compliance_passed, summary
Execution harness: Python
Evaluation tools
- CrewAI · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.