A stack of construction subcontractor agreements on a job site trailer desk, with a laptop screen showing a redlined contract document, warm afternoon light through a dusty window
Policy & Regulation

Utah Gave an AI a Law License. It Charges $117 to Review the Contracts Your GC Never Reads.

By Catherine Chen · June 4, 2026

David Fitzhugh manages contract strategy at Western Partitions, an interior specialty contractor in Portland, and he used to send every subcontract and supplier agreement to outside counsel for review at rates that made the line item painful to look at. "We are probably spending a tenth of the cost that we did with outside legal," Fitzhugh told TechStartups, "and 85 to 90 percent of contracts are turned around in 24 hours. I've never waited more than two days for a contract review. I cannot say the same thing about outside legal counsel."

Superlegal launched June 3, calling itself the first AI law firm in the United States authorized to practice law, operating under the Utah Supreme Court's Legal Services Innovation Sandbox, which is a regulatory experiment allowing approved entities to deliver legal services through technology-driven models and non-lawyer ownership structures that would be unauthorized in every other state. Price: $117 per contract. A licensed attorney signs off on every review.

That last sentence is doing a lot of work.

What $117 Actually Buys

Superlegal's AI ingests a commercial contract, identifies risk clauses, flags deviations from standard terms, and produces a redlined version with annotations explaining what each deviation means, why it matters, and what alternative language to consider. Then a licensed attorney reviews. Twenty-four hours. Upload to redline. Superlegal claims 90 percent cost reduction and 70 percent shorter deal cycles, with construction firms including BGE, Silverline Construction, SHN, Glass, Great Lakes Construction, and SSI Aeration already on the platform.

Founded by Noory Bechor and Ilan Admon, who previously built LawGeex (an enterprise contract review platform acquired in 2023), Superlegal raised $5 million in seed funding in 2024 from Aleph, Disruptive AI, former Thomson Reuters CEO Tom Glocer, and the Google AI Startups Fund. It has also partnered with the Associated General Contractors of America, giving AGC members direct access to the platform.

"Most legal AI companies sell tools to lawyers, and the lawyers still bill clients $500 an hour," Bechor said. "We built our AI from the ground up to specialize in legal practice, with an attorney in the loop. The Utah license means a construction firm can hire us, full stop."

$17,660
Estimated savings per residential project if a GC reviews 20 subcontractor agreements through Superlegal ($2,340) instead of traditional counsel (~$20,000 at $1,000/contract average). That's 1 to 3.5 percent of total project cost on a $500K to $2M custom home.

Why Construction, Why Now

According to AGC survey data, 61 percent of construction firms are now using or planning AI investments, up from 44 percent in 2024. Companies deploy AI for office functions (45 percent), estimating (23 percent), and design (20 percent), with 38 percent reporting measurable business impact according to ServiceTitan. Contract review is next.

Consider the math on a custom home, where a typical general contractor managing a $500,000 to $2 million residential project signs between 15 and 30 subcontractor and supplier agreements, each containing scope definitions, payment terms, liability allocations, insurance requirements, dispute resolution provisions, and lien waiver language that varies by state, adding up to a thicket of legal exposure that most small builders never fully map before signing. Thorough review by a construction attorney takes two to three hours at $450 to $500 per hour, running $900 to $1,500 per contract, and if you multiply by 20 you hit $20,000 in legal fees on a single residential project.

Most small contractors skip that entirely and sign blind. When disputes arise, the average construction dispute in North America reached $42.8 million in value by the end of 2022, according to Arcadis's annual construction dispute report, with the leading cause being parties failing to understand or comply with their contractual obligations. That figure reflects commercial-scale disputes, not residential ones, but the mechanism is identical.

Nobody Mentions Jurisdiction

Superlegal is authorized to practice law in Utah and nowhere else. Construction contracts, however, are governed by the law of wherever the project sits, which means a general contractor building homes in Texas needs a reviewer who knows Texas contract law, Texas mechanic's lien statutes, Texas prompt payment provisions, and whatever municipality-specific bonding requirements apply, all before touching the indemnification clause. An AI trained on general commercial contract language may flag a problematic indemnification provision, but miss the distinction between a Texas pay-if-paid clause (enforceable as a condition precedent under Texas Business & Commerce Code Chapter 56, but defeasible by subcontractors who follow statutory notice procedures) and a pay-when-paid clause (treated as a timing mechanism, not a condition precedent). That distinction has destroyed subcontractors.

State-specific construction law is dense, technical, and hostile to generalization. Mechanic's lien filing deadlines range from 60 days in some states to 180 in others, preliminary notice requirements differ radically, and retainage statutes cap withholding at different percentages depending on project type and jurisdiction. Pattern-matching against a corpus of commercial contracts does not get you there.

Superlegal has not disclosed what jurisdictional coverage its AI provides, how its model handles state-specific provisions, or what liability framework applies when a reviewed contract leads to a dispute the review should have caught. For a contractor whose entire business rides on getting the lien waiver language right, that silence is not a minor omission.

A Shrinking Sandbox

Utah's Legal Services Innovation Sandbox is itself an experiment with a shrinking population. A Stanford Law School study tracking the first five years of regulatory reform found participation contracted sharply: 39 entities down to 11. Twenty-seven out. Rocket Lawyer, one of the original entrants, left for Arizona, where a more permissive alternative business structure regime has grown from 19 to 136 licensed entities.

Utah now requires sandbox participants to demonstrate they reach underserved consumers, and the entire program expires in August 2027. Only 20 consumer complaints total have been filed against sandbox entities as of April 2025, which Stanford characterizes as evidence that consumer harm remains minimal. But minimal harm is not proven value, and a regulatory experiment with a fixed expiration date is not the foundation most contractors expect from their legal counsel, especially when the entire authorization to practice could vanish in thirteen months.

What $117 Buys and What It Doesn't

A licensed attorney signs off on every Superlegal review, but at $117 per contract, think about what that sign-off can realistically involve. If the attorney spends 15 minutes, they earn $468 per hour before AI costs and Superlegal's margin, which is competitive with BigLaw rates and assumes the AI has done the substantive analysis correctly so the attorney is confirming rather than re-reviewing from scratch, which at that price point is the only way the economics work for anyone. If they spend five minutes, the sign-off becomes a rubber stamp, and the malpractice exposure calculus changes entirely.

Construction punishes missed clauses brutally. One overlooked pay-if-paid provision or one overbroad indemnification carve-out can shift $200,000 in liability overnight. Traditional construction attorneys carry errors-and-omissions insurance sized to those stakes, paid for by fees that justify thorough review. Superlegal has not disclosed its malpractice coverage, the average attorney review time per contract, or the error rate of its AI on construction-specific provisions.

If You Are Considering This

For a GC running $500,000 to $2 million residential projects with 15 to 30 sub agreements per job: start small. Test Superlegal on commodity material supply agreements with standard terms and limited liability exposure. Compare its redlines against your attorney's on the same document. See where they diverge. Use the savings on boilerplate to fund deeper attorney engagement on the contracts that actually matter, which means your prime contract with the owner, your high-value sub agreements, and anything with indemnification provisions that could survive project completion and haunt you two years after the certificate of occupancy.

Do not use any AI-powered review as sole legal analysis on contracts involving mechanic's lien waivers, conditional payment clauses, or performance bond obligations. These provisions interact with state-specific statutes in ways that demand human judgment.

If Superlegal survives the sandbox expiration and builds real jurisdictional depth, it could change how small contractors manage legal risk. Right now it is a promising tool operating under a temporary license in a single state, with a pricing model that raises more questions about attorney oversight than it answers. Construction desperately needs cheaper contract review. Whether $117 delivers that, or merely the appearance of it, is a question no one has answered yet.

Limitations

All cost savings figures cited in this article originate from Superlegal or its customers and have not been independently audited. A per-project savings estimate of $17,660 assumes 20 contracts per residential project at average traditional review rates of $1,000 per contract, based on published attorney billing surveys for construction law, and actual costs vary significantly by market and firm size. Arcadis dispute data reflects commercial-scale construction, not residential projects, and is used here to illustrate the contract-dispute mechanism rather than to quantify residential exposure directly. Superlegal's customer list as published includes no residential-focused contractors, meaning the platform's applicability to custom home construction is projected, not demonstrated. Utah's sandbox entity count may differ from the Stanford study's April 2025 data. Attorney time estimates for the $117 review are analytical inferences, not disclosed figures from Superlegal.

← Back to AI Home Building