LawGeex

What it is

LawGeex stopped existing as a buyable product in 2023. Robin AI acquired the assets of LawGeex’s enterprise division on February 27, 2023, and LegalSifter acquired a strategic group of the remaining enterprise clients — dozens of companies, 17 of them Global 2000 — on September 25, 2023. Nothing continued under the LawGeex name. The domain redirects HTTP to HTTPS and then fails on an expired TLS certificate (checked 2026-07-27), so lawgeex.com is unreachable in a normal browser. This page stays up for historical context and to route former customers, and anyone searching for what happened to LawGeex, to the products that now do the job.

LawGeex was a pre-signature contract-review platform founded in Tel Aviv in 2014 and funded to roughly $45M across four rounds, ending with a $20M Series C in April 2020. It took inbound counterparty paper — NDAs, vendor agreements, DPAs — compared each document against the buyer’s playbook, and either auto-approved it or returned a redline for a lawyer to finish. Its best-known artifact was a February 2018 report pitting the algorithm against 20 US-trained lawyers on NDA issue-spotting, scoring 94% for the software against an 85% average for the lawyers. Academics and an independent consultant helped design and oversee it, but LawGeex published it itself; it was not peer-reviewed in a journal, and the spread among the human participants was wide — the best lawyer also hit 94%, the weakest 67%.

The 2022 restructuring is where the ending starts. LawGeex cut roughly a third of its staff and split in two: the enterprise platform, and an SMB-focused venture called Superlegal. The enterprise side never recovered, and within eighteen months it had been sold twice and dismantled.

Where the pieces sit now: the asset trail dead-ends. Robin AI, which bought the technology, collapsed itself after a funding round failed to close and an HMRC winding-up petition landed — it sold its managed-services arm to Scissero in December 2025, and Microsoft acqui-hired its engineering team in January 2026 to work on Word. LegalSifter, which bought the client relationships, is the one live destination: it still sells ReviewPro for AI redlining in Word and Contract Logix for lifecycle management. Founders Noory Bechor and Ilan Admon took the Superlegal side, raised a $5M seed in 2024, and relaunched it on June 3, 2026 as an AI law firm for US construction, authorized under the Utah Supreme Court’s Legal Services Innovation Sandbox, reviewing and redlining contracts from $117 each in under 24 hours.

Why it showed up in Legal Ops stacks

• It was built for inbound paper. Most CLMs assumed your team drafted the contract from a template. LawGeex inverted that: the counterparty sent their paper, the engine read it against your playbook, and it decided what was acceptable. That framing is now the default for the whole contract-AI category, which is LawGeex’s real legacy.
• Auto-approval, not just redlining. Configurable thresholds let routine NDAs and vendor contracts inside playbook tolerance clear without a lawyer opening them. The share of inbound paper that cleared was vendor-reported and never independently audited.
• Rule-traceable decisions. Every clause verdict pointed at the playbook rule it matched. Legal Ops could audit an auto-approval back to the rule that drove it — a stronger audit story than the generative tools that replaced it, and the thing former customers miss most.

Pricing reality (historical)

LawGeex never published a price. Deals were quoted on contract volume and playbook complexity rather than per seat, and the sales motion ran through a 60-90 day pilot on a single contract type, almost always NDAs. Every dollar figure still circulating for LawGeex traces to competitor pricing pages and reseller breakdowns rather than to the vendor, and with the company gone none of it is verifiable against a source that ever had the numbers. Treat any LawGeex price you find in a 2026 comparison article as unsourced.

What it was best for

In-house legal and procurement teams whose bottleneck was review throughput on repetitive inbound paper — sales-led organizations drowning in counterparty NDAs, procurement groups reading vendor terms at volume, biotech BD teams. It fit organizations large enough to justify four to eight weeks of encoding their standards into a rules engine, and it was the wrong buy for teams that mostly drafted their own paper or signed too few contracts to amortize that setup.

Versus the alternatives (now that LawGeex is gone)

LegalOn is the closest replacement for the whole LawGeex motion: playbook-driven review of counterparty drafts, 50+ attorney-written standards shipped with the product, and a Triage Agent that auto-approves low-risk agreements or escalates them — the direct descendant of LawGeex’s auto-approval thresholds, without the multi-week rules encoding. BlackBoiler is the pick when the job is narrower and higher-volume: automated markup of third-party paper against your positions, with no ambition to be a review workspace. Luminance fits teams that want the counterparty negotiation itself automated on high-volume agreement types. Ivo matches on the in-house review-and-playbook job when your agreements sit in one jurisdiction. Spellbook publishes a $99 floor and is the drafting-led choice, weaker on inbound triage but the only one on this list you can price without a sales call. If the constraint is a CLM rather than a reviewer, Ironclad and Agiloft own the lifecycle side.

Two destinations sit outside the ooligo catalog and are worth naming anyway. LegalSifter is where LawGeex’s enterprise clients actually landed, and its ReviewPro runs the same first-pass-review job inside Word. Superlegal is the founders’ current venture and sells the outcome instead of the software — a reviewed, redlined contract at a per-document price, with a licensed attorney signing off — which suits teams that wanted LawGeex’s result without owning a playbook.

Watch-outs (for teams evaluating replacements)

• Vendor-published accuracy benchmarks are the category’s founding habit, and LawGeex set it. The 94% figure was measured on a curated NDA sample by the company selling the software, and every contract-AI vendor since has published a number in the same shape. Guard: run a paid pilot on 30-50 of your own live agreements, have a senior reviewer grade the output against what they would have written, and gate the contract on the miss rate you measured rather than the one on the slide.
• Playbook encoding is the real cost, and it does not survive the vendor. LawGeex customers spent four to eight weeks translating their standards into a rules engine; when the product was dismantled, that work stayed inside a platform that stopped running. Guard: before signing with any replacement, require a contractual right to export your playbook rules and clause positions in a readable format on demand, not only at termination, and test the export once during the pilot.
• This corner of legal tech changes hands faster than a contract term. LawGeex, Robin AI, and Casetext all disappeared into other companies inside three years, and two of the three took their platforms down with them. Guard: ask any shortlisted vendor for its funding position and runway in writing, keep your executed contracts and extracted data in a system you control rather than in the review tool, and write a data-return-on-termination clause with a fixed deadline into the order form.
• A dead tool’s page keeps ranking, and stale comparisons keep recommending it. LawGeex still appears as a live option in third-party 2026 buyer’s guides and pricing roundups. Guard: before shortlisting any contract-AI vendor from a listicle, load the vendor’s own site and check that a demo request or pricing page actually resolves — the check that would have caught this one takes ten seconds.