Uber’s Legal AI Agent Cuts Contract Review Time by 20% with Rules and RAG

InfoQ (AI, ML & Data)
Read full story· 3 min
Uber's Legal Redlining Agent, integrated with Microsoft Word, cuts contract review time by over 20% with 91% accuracy by combining retrieval-augmented generation, lawyer feedback, and rules. It analyzes contract edits, assesses risks, and suggests policy-aligned responses while lawyers retain final approval.

What happened

Uber developed and deployed a Legal Redlining Agent that works within Microsoft Word to help its legal team review contracts. The tool uses retrieval-augmented generation, rules, and lawyer feedback to analyze contract edits, assess risks, and suggest responses aligned with company policy. Lawyers remain responsible for reviewing and approving the AI's recommendations. The system was first developed in early 2024 and piloted with Uber's legal team in 2025.

Key details

  • The agent reduced contract review time by over 20%.
  • It achieved 91% accuracy in its AI-generated decisions.
  • The tool integrates with Microsoft Word as an add-in.
  • It uses a combination of retrieval-augmented generation, deterministic rules, and historical lawyer feedback.
  • A rules engine enforces non-negotiable legal requirements and can override model suggestions.

Why it matters

Uber's legal team can now process contracts more quickly, saving time on repetitive review tasks. Lawyers maintain control over final decisions, ensuring compliance and oversight. The system offers a concrete reduction in workload and turnaround time for contract review, which can impact operational efficiency and legal risk management for the company.


More in Business & Enterprise