4
min. read

How long does AI contract review take?

Jeff Dutton
By
Jeff Dutton
Lawyer
Last update:
September 4, 2026

Review any Contract With AI Before you Sign it

A few minutes. Point an AI first pass at a normal commercial agreement and you wait a few minutes while it reads the document, breaks it into clauses, and hands back a list of findings. Your contract still takes days or weeks to get signed. The reading was never the slow part.

Those are two different clocks, and the gap between them is where this question usually goes wrong.

The clock you are actually measured on

APQC, the benchmarking non-profit, tracks a measure called average cycle time in days to establish a contract with a supplier, counted from the day a negotiation opens to the day the thing is signed. Across a sample of 3,081 companies the median is 40 days.

The definition matters more than the number. APQC counts cycle time as “both time spent actually performing the process and time spent waiting to move forward.” Waiting is inside the measurement, on purpose.

So when somebody asks how long AI contract review takes, they usually mean one of two very different things: how long the software runs, or how long a contract now takes to clear your team. The first answer is minutes; the second depends on your queue.

What the few minutes are doing

A review is not one call to a model. In goHeather it runs in three stages the app names while you watch: splitting the contract into clauses, reviewing those clauses for risks and obligations, then generating the summary. Clause detection has to come first, because every finding after it is anchored to a real section number, and you cannot anchor anything in a second and a half.

Setup sits inside that time too. Clicking review asks you nothing. It reads the document, works out both parties, preselects your side, pulls the governing law off the page, and puts all of it in front of you as editable settings before anything runs. That is the orienting a reviewer would otherwise do before reading a word of the agreement, and it is a decent chunk of what an AI first pass on an inbound contract is doing while you sit there thinking it is being slow.

Long documents take longer, and the app says so rather than pretending otherwise: this takes a few minutes, and you can keep chatting once it lands. The work runs server side, so you can close the tab.

The wait nobody quotes

Here is the part that actually decides your turnaround. A review comes back with findings, and each one sits there until a person accepts it or dismisses it. Progress is tracked as a count and a percentage, and the chat is gated on it: nine review findings still need a decision, accept or dismiss to keep chatting.

Counts in our own account run from one to twenty-five findings on a single document.

The software does not close a finding. A person does, and the product will not move past them until somebody has. If the reviewer who owns that file is in back-to-backs until Thursday, the review took two days. The minutes are not what you were waiting on.

Why the second-count claims miss

You will see AI review times quoted in seconds, and turnaround improvements quoted as a percentage cut. Both are measuring the read.

Quality engineers have a name for the gap. The value-added ratio is the share of total elapsed cycle time spent actually working on something, as opposed to the thing sitting in a pile. Writing in ASQ’s Quality Progress in 1997, Patrick Shannon of Boise State put the typical ratio under 10 percent, and often under 1. That is a general finding about business processes, not a measurement of contract review. It still sets the expectation honestly. If the reading is a small share of forty days, taking the reading to nearly zero leaves most of the forty days sitting exactly where it was.

Which is the same reason a review backlog keeps growing even when your reviewers get faster.

So, the honest answer. The software takes minutes. Your contract takes as long as your slowest decision, and the value of a first pass is that it puts those decisions in front of a named person as a list, with the clause text attached, instead of leaving them buried in a file nobody has opened yet.

Try goHeather free and time your own first pass against your own paper. This is legal information, not legal advice; consult a lawyer for legal advice.

About the author

Jeff Dutton is a lawyer who advises on technology, corporate, privacy, commercial, employment and real estate law.

Jeff founded his own small law firm, Dutton Law, in 2016 (and merged it with a larger firm in 2019). Before that, Jeff was a prosecutor and a commercial law lawyer at a national boutique law firm.

Jeffrey is a frequent lecturer on legal matters and has been published in newspapers and trade journals. In addition, Jeff was the editor and co-author of a leading employment law text for lawyers for many years.

Education:

Western University, BA (2009)
University of Ottawa, Faculty of Law, JD (2012)

Jeff Dutton
By
Jeff Dutton
Lawyer

Stay Updated on All Things Contract Law with goHeather

Get the latest contract tips, updates, and exclusive content straight to your inbox. Subscribe now and never miss out on what's new in contract law or at goHeather!

Thank you! You will receive an email to confirm your subscription.
Oops! Something went wrong while submitting the form. Try again later.

Review any Contract With AI Before you Sign it

Our AI sifts through each clause, identifying potential risks. This enables us to provide quick yet comprehensive contract reviews, equipping you with the legal information you need to make informed decisions.

Related articles