Contract Review Time Benchmarks: How Long Does Contract Review Actually Take?

"How long does contract review take?" has surprisingly few honest, sourced answers - most numbers in circulation are vendor folklore, detached from whatever study once produced them. This page compiles the published benchmarks on contract-review time and cost, each cited to its original source, with the caveats that the original studies themselves attach. It is a companion to our legal AI statistics page; where that page covers adoption and ROI broadly, this one covers a single question: the time and cost of reviewing contracts, manually and with AI. Figures verified July 2026; refreshed quarterly.

How long does contract review actually take?

The short, sourced answer: a trained lawyer needs roughly 15-20 minutes of reading time for a short, standard agreement (the LawGeex study's lawyers averaged ~18 minutes per NDA), while organisational turnaround for the same document is typically measured in days - because review time and queue time are different things. Published company baselines back the distinction: Trench Group reported an average of about 150 minutes of review effort per contract before automation, and Demandbase reported NDA turnaround of 2-3 days before moving first-pass review to AI.

Every number that follows is one of three kinds - controlled-study time (minutes of reading), organisational cost (dollars per processed contract), or cycle time (days from request to signature) - and conflating them is how folklore statistics get born. The tables below keep them separate.

The LawGeex study: 26 seconds vs 92 minutes

The most-cited controlled experiment in this field is the February 2018 LawGeex study, which put an AI trained on NDAs against twenty US-trained corporate lawyers (alumni of Goldman Sachs, Cisco, and major firms among them) on the same task: issue-spotting five real NDAs from the Enron dataset - 153 paragraphs, ~30 provision types, four-hour limit.

Three caveats belong to the citation. The 92 minutes covers all five NDAs - quoting it as per-contract time overstates human effort five-fold. The task was issue-spotting against a defined checklist, not negotiation or judgment. And the best individual lawyer matched the AI's 94% - the study's fair summary is not "AI beats lawyers" but "AI delivers top-quartile issue-spotting consistency at effectively zero marginal time".

What a contract costs to process: the WorldCC benchmarks

The average cost of processing a basic everyday contract is $6,900, per the benchmark study by IACCM - now World Commerce & Contracting - published November 2017 from analysis of 700+ organisations, and still the field's reference point. The full tiering:

A widely repeated "$700 per simple contract" figure could not be traced to any named study - treat it as folklore. WorldCC's related research is just as quotable and better sourced: its 2023 study with Deloitte (1,236 organisations) put the value leaked through poor contracting at 8.6% of contract value on average (best performers ~3%), and its Most Negotiated Terms 2024 report - the study's 22nd year - found limitation of liability still the single most-negotiated term, a crown it has held for over a decade. For clause-level review practice, our essential clauses checklist maps directly onto that most-negotiated list.

Cycle times and business impact: the EY-Harvard data

The largest study of contracting-process pain is the 2021 EY Law / Harvard Law School Center on the Legal Profession survey of 1,000 contracting professionals. Its headline findings, still the standard citations for contracting friction:

• More than 50% of organisations said contracting inefficiencies had slowed revenue recognition; 50% reported actually losing business because of them; 57% of business-development leaders reported slower revenue.
• 92% were transforming their contracting process - yet 99% lacked the data and technology to do it well, and 38% had already tried and failed.
• 65% of contracting staff spent their time on unchallenging, low-complexity contracts; over 40% of contracting time and budget went to that low-value work; 89% of companies called high volumes of low-complexity contracts a departmental challenge.

The companion EY-Harvard general-counsel research (April 2021) frames the capacity squeeze those numbers create: GCs expected workloads to rise 25% over three years with headcount growth of only 3%, and 87% said their department spent too much time on low-value routine work. Contract review is where that squeeze concentrates - the same routine paper, at rising volume, against flat teams.

Volume: how many contracts teams actually handle

Volume benchmarks put the time-per-contract figures in context. The EY-Harvard study found large organisations manage an average of 350 contracts per week - roughly 19,000 a year, with the busiest handling 50,000+. The most famous single-company datapoint remains JPMorgan's COIN programme: contract-intelligence software that took over the interpretation of 12,000 commercial credit agreements a year, work that had consumed an estimated 360,000 hours annually of lawyer and loan-officer time (Bloomberg, 2017 - note the "and loan officers": the hours were never lawyers alone).

Multiply the WorldCC per-contract averages by these volumes and the arithmetic explains why contracting became the first legal workflow to industrialise: 19,000 contracts a year at even the top-quartile simple-contract cost is a very large number before a single complex deal is counted.

AI-assisted review: the published before/after numbers

Published before/after figures for AI-assisted review - with the honest label that most are vendor-published or vendor-commissioned, though they name real companies:

The Ashurst trials deserve the extra weight of being a law firm publishing its own measured results: 80% time savings drafting UK corporate filings, 59% on research-report drafting, about 45% - roughly 2.5 hours saved - per first-draft legal briefing. The pattern across all five rows is consistent: the first pass compresses dramatically; the lawyer's judgment layer remains. That is the same design principle behind Judicio's Document Review, where AI runs the clause-by-clause first pass with page-cited findings and the reviewing lawyer works from a severity-ranked report rather than a blank contract.

What actually drives review time

Across the studies, four variables explain most of the variance in review time - worth knowing because they are also the levers. Complexity tier: WorldCC's 7x cost spread between basic and complex contracts is the dominant factor. Familiarity: review against a known playbook (your positions, your fallbacks) is systematically faster than open-ended reading - the mechanics are covered in our playbook review guide. Queue position: the days-long turnarounds in the case studies were queues, not reading time - which is why intake automation moves the business metric even when reading time is unchanged. Second-pass burden: rework after a missed issue costs more than the first pass; the LawGeex accuracy gap (85% average, 67% floor) is really a rework statistic in disguise.

How to use these benchmarks on your own workload

Benchmarks corroborate; they do not substitute for your own baseline. The practical sequence: measure two weeks of your team's contract flow (volume by complexity tier, minutes per review, queue time from request to return), place your numbers against the tables above, and target the biggest gap - usually the low-complexity, high-volume tier that the EY-Harvard data shows eating 40% of contracting capacity. Our business-case guide turns that exercise into a budget memo, and the review-speed guide covers the workflow changes that move the numbers. To test the AI-assisted side of the table on your own paper, start a free 7-day trial - 500 credits, no card required - and run last month's NDAs through a playbook review.

Sources and methodology

Primary sources, linked in place: LawGeex, Comparing the Performance of Artificial Intelligence to Human Lawyers in the Review of Standard Business Contracts (Feb 2018); IACCM/WorldCC, The Cost of a Contract (Nov 2017, 700+ organisations) and WorldCC-Deloitte, The ROI of Contracting Excellence (2023, 1,236 organisations); WorldCC Most Negotiated Terms 2024; EY Law / Harvard Law School CLP contracting survey (2021, n=1,000) and GC survey (2021); Bloomberg on JPMorgan COIN (2017); Ashurst Vox PopulAI (2024); named vendor case studies from Luminance and Ironclad and the Forrester TEI of DocuSign CLM - each labelled as vendor-published or commissioned where applicable.

Methodology notes: we distinguish reading time, processing cost, and cycle time throughout; vendor-published figures are included only where the customer is named, and are labelled; the "$700 simple contract" figure circulating in vendor content could not be traced to any named study and is excluded. Figures verified July 2026; this page is refreshed quarterly and anything older than 18 months without a stable primary source is re-verified or removed per our editorial standards.