The Document Review Bottleneck at Second Request Scale
Four million documents is not unusual for a major DOJ or FTC second request — what's unusual is having 30 days to process them. At conventional review rates, even a large contract review team cannot process that volume, code for responsiveness and privilege, and generate a defensible privilege log in the available window. The privilege log problem is particularly acute: a log assembled under time pressure from manual reviewer notes is the most likely target for a government challenge, and a successful challenge creates discovery exposure that extends far beyond the original production. The review technology most firms have deployed can assist reviewers; it cannot run the workflow.
A Workflow That Runs Itself, With Human Review at the Edges
An AI Labor Company agent deploys a continuously-updated responsiveness and privilege coding workflow directly in Relativity. The agent processes documents against your responsiveness criteria and privilege taxonomy, making coding decisions on clear-cut documents and routing edge cases — ambiguous privilege calls, multi-party communications, and documents with responsive and non-responsive content — to associate review queues in Relativity. Privilege log entries are generated in Concordance-ready format as documents are coded, rather than assembled after the review is complete. The agent can also cross-reference Westlaw Edge for privilege doctrine questions that affect coding logic in specific jurisdictions.
Billing Capacity and Matter Competitiveness
The business case for second request automation is about matter economics and firm capacity. A second request at $250K–$600K in projected billings has a cost structure dominated by document review — and a review process that runs faster and produces a more defensible privilege log is a competitive differentiator when clients are evaluating firms for major antitrust matters. Firms that can credibly propose a 30-day second request response with a defensible privilege log, rather than a 60-day response contingent on staffing up a large contract review team, win mandates that others cannot take. The efficiency reduction in review time is typically 65–85%; the privilege log is a byproduct of the review, not a downstream assembly project.
How does the agent's privilege coding hold up to a government challenge?
The agent applies your firm's documented privilege taxonomy consistently across all documents, generating a log entry with the underlying coding rationale for each privileged document. Consistency and documented rationale are the two factors most likely to survive a government challenge — both are stronger in an agent-run workflow than in a manual review under time pressure.
Can the agent handle foreign-language documents in a cross-border second request?
The agent can be configured for multi-language document sets. Foreign-language documents can be routed to a translation-first queue or processed against translated content, depending on the volume and language distribution of the document set.
What happens when the government requests a supplemental production mid-review?
Supplemental production requests are handled by re-running the relevance criteria against the document universe with updated parameters. Because the coding workflow is continuous rather than batch-based, the agent can incorporate revised criteria without restarting the entire review.