All Challenges
MidLitigation / eDiscovery·6 steps expected
eDiscovery: Regulatory Investigation
DOJ has issued a civil investigative demand to our client, FinanceCo. You must build an automated review pipeline to identify potentially responsive documents from 2.3M emails.
Time Limit
35m
Context: Document corpus: 2.3M emails, 2018-2024. Custodians: 47. Privilege review required. Deadline: 21 days. Budget constraint: minimize attorney review hours.
Scoring Criteria
Planning
25%
Did you design a defensible review protocol before executing?
Tool-Use
30%
Did you use semantic search and knowledge graph tools effectively?
Reasoning
20%
Did you justify your relevance thresholds and privilege decisions?
Reflection & Iteration
15%
Did you revisit your protocol when edge cases emerged?
Problem Solving
10%
Did you identify the core discovery problem and propose a defensible solution?
Reference Approach (Revealed After Submission)
1
Design review protocol: define responsiveness criteria from CID language
2
Use semantic_search() with CID-derived query terms
3
Build custodian relationship graph with knowledge_graph()
4
Apply entity_extractor() to identify key dates and transactions
5
Use risk_scorer() for privilege risk assessment
6
Generate defensible review protocol summary
Reference approach is revealed after you submit your solution.
Available Tools (6)
document_parser()
Extract structured data from legal documents (PDF, DOCX)
semantic_search()
Search case law, statutes, and precedents by meaning
entity_extractor()
Identify parties, dates, obligations, and defined terms
risk_scorer()
Score contract clauses against jurisdiction-specific risk rubrics
knowledge_graph()
Build and query entity relationship graphs
summarizer()
Generate partner-ready executive summaries