The Pulse
DOJ Seeks Information From Vendors for 20- to 30-User Legal AI Pilot
Nextgov/FCW reported that the Justice Department is asking vendors about commercially available legal AI platforms for a potential pilot involving 20 to 30 government users. No named speaker appears in the source; the quotations are from th
AI.info Team ·
The Justice Department is asking vendors for information about commercially available artificial intelligence legal-assistant platforms that can analyze documents and provide legal research, according to Nextgov/FCW, which cited a SAM.gov document posted Friday, September 18. The information could shape a pilot involving 20 to 30 government users.
The proposed pilot would test the technology in a vendor’s FedRAMP-authorized or commercially available cloud environment using synthetic or public data. The request is for market research and does not launch a procurement or guarantee a contract.
Nextgov/FCW reported the request on September 21. The publication said a DOJ spokesperson did not respond to questions about why the department is interested in an AI legal-assistant platform or when it might pilot the technology.
DOJ wants an assistant that can work across legal tasks
The department is not limiting the proposed evaluation to question-answering or document search. The platform capabilities described in the request include document analysis, regulatory review, contract clause extraction, Freedom of Information Act processing and memo drafting.
“autonomously execute multi-step legal tasks, including document analysis, regulatory review, contract clause extraction, [Freedom of Information Act] processing, and memo drafting — without requiring step-by-step attorney prompting.”
The proposed system would therefore be evaluated on its ability to handle multiple stages of legal work without attorneys providing instructions at every step. The document does not identify a preferred model, vendor or technical architecture.
DOJ also plans to assess whether a platform can synthesize between 5,000 and 100,000 documents per session, provide authoritative sources for legal research, assist attorneys with drafting documents and integrate with existing agency document-management systems.
A controlled test using public or synthetic data
The pilot would be conducted in a vendor’s FedRAMP-authorized cloud environment or another commercially available cloud setting. Testing would use synthetic or public data rather than live sensitive case material.
Evaluation would combine technical and user-based measures. DOJ plans to consider platform-provided usage statistics, government-administered reviews of output quality, contractor-provided timing logs, technical demonstrations, government checklist reviews and anonymous employee feedback surveys.
The request builds on DOJ’s existing legal software
DOJ has already deployed AI tools from LexisNexis, Westlaw and Bloomberg to assist with legal research, according to the department’s 2025 AI use case inventory, as reported by Nextgov/FCW.
The proposed assistant would be tested alongside tools already used for legal research, but the request describes a broader set of capabilities. Those include processing large document collections, producing drafts and handling several steps in a legal workflow.
The document is not a contract award
The SAM.gov document describes the request as market research rather than a formal solicitation.
“This is a market research exercise only — responses will not be evaluated as offers and will not result in a contract award. Responses inform the agency's acquisition planning.”
The language leaves open whether DOJ will issue a formal solicitation, how it will handle attorney review of AI-generated work and what standards will govern source accuracy. The current request establishes a test concept and a list of desired functions, not a commitment to deploy an autonomous legal system.
For now, the department is seeking information about whether existing platforms can handle legal work at government scale while operating in an approved cloud environment and producing measurable evidence about speed, quality and employee use.