M&A Due Diligence Compliance Analyzer with Mastra AI
Financial institutions conducting bank M&A and private equity acquisitions struggle to audit acquired portfolio companies for hidden regulatory and cybersecurity risks. Build an automated due diligence agent using Mastra AI in TypeScript to ingest data rooms, audit compliance documentation, and flag unmitigated vulnerabilities. Achieve 90%+ recall on hidden regulatory non-compliance findings.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an automated M&A compliance audit workflow using Mastra AI in TypeScript to inspect virtual data room filings and detect regulatory risk exposures.
How work is evaluated
Evaluates due diligence audit accuracy and completeness of flagged compliance deficiencies.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
unencrypted_pii_flag_test
Verifies that unencrypted storage finding is caught and results in HIGH risk classification
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
finding_recall
Recall score of ground-truth non-compliance issues identified • target: 0.9 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Configure Mastra AI Agent workflows with TypeScript schema definitions
Parse multi-document virtual data room repositories containing audit records
Build compliance rule evaluation nodes for detecting unmitigated AML/SOC2 deficiencies
Generate actionable audit executive summary reports with structured risk metrics
Reference links and supporting material
Dataset of 50 virtual data room document bundles containing disguised regulatory audit findings and compliance logs.
Challenge setup
Evaluates due diligence audit accuracy and completeness of flagged compliance deficiencies.
Challenge input
JSON data room object with document titles, text chunks, and category labels
Mastra AI
TypeScript-native agent framework
Evaluated output
JSON summary with aggregate_risk, non_compliance_findings array, and pass_recommendation
- Verifies that unencrypted storage finding is caught and results in HIGH risk classification
- Recall score of ground-truth non-compliance issues identified • target: 0.9 • range: 0-1
- Benchmark: AuditRisk-Eval
- Finding Recall target: 0.9
- 1 public reference case
- JavaScript execution harness
- javascript 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
- Mastra AI · Required
- Mastra AI · Required
Evaluation contract
- unencrypted_pii_flag_test · Weight 1
- finding_recall · Weight 1
Metrics
- finding_recall · Target 0.9
Benchmarks
- AuditRisk-Eval
- security
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON data room object with document titles, text chunks, and category labels
Output: JSON summary with aggregate_risk, non_compliance_findings array, and pass_recommendation
Execution harness: JavaScript
Evaluation tools
- Mastra AI · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.