Skip to content
Trending storyDeveloping

Uber's legal team cuts contract review time 20% with AI redlining add-in

1 article1 sourcesince Oct 10Last article 2h ago ·

Overview

AISummary of 1 article

Uber's legal team built a Microsoft Word add-in that suggests accept, reject, or modify decisions on contract redlines, with lawyers making the final call.

Per Uber Engineering, the team moved from plain retrieval-augmented generation (RAG) to a feedback-driven system with agentic drafting and a rules database, raising AI decision accuracy to 91% and cutting average contract review time by more than 20%. The source reports these results in four iterations of redesign, as summarized by a post on X from @shao__meng.

Written by AI from the articles below · updated Oct 10, 4:54 AM ET

Check the sources:

Latest developmentOct 10, 4:23 AM ET
Uber's legal redlining agent moved from failed RAG to 91% accuracy in four iterations

Article timeline

The articles in this story. Times are ET.

Oct 10
  1. meng shaoX
    Uber's legal redlining agent moved from failed RAG to 91% accuracy in four iterations

    AIUber's legal team built a Microsoft Word add-in that suggests accept, reject, or modify decisions on contract redlines, with lawyers making the final call. Per Uber Engineering, the redesign from plain RAG to a feedback-driven system with agentic drafting and a rules database raised AI decision accuracy to 91% and cut average contract review time by over 20%.

    Image from @shao__meng's post

Heat trend

Not enough continuous observations to show a trend yet.