Orange Collective
General Legal

General Legal

The AI-native law firm for growth-stage companies.

general.legalYYC Profile@gen_legal_incLinkedIn
General Legal Launch[3]

$500

Per contract

All turns included · flat fee

<3 hrs

Turnaround SLA

Drafting, review, negotiation

20+

Growth-stage logos

Signed since private preview

Thesis

Commercial contracting is on the revenue critical path — and outside counsel is the most expensive, slowest, and least predictable part of it.[5] General Legal is the AI-native law firm that prices per contract instead of per hour, turns drafts in under three hours, and pairs proprietary AI workflows with experienced attorneys who own the work product. The wedge is fast, flat-fee contracts for growth-stage tech. The long arc is the default commercial contracting desk for every AI-era startup — and a compounding "what's market" dataset that no incumbent firm or tool-only competitor can replicate.[4]
  1. 01

    Flat-fee, outcome-aligned pricing is the structural unlock. Hourly billing misaligns incentives for routine contracting — the firm gets paid more when the work takes longer. Per-document pricing focuses operators on throughput, quality, and deal velocity. $500 per contract (all turns included) with a sub-3-hour SLA hits the exact pain founders feel from outside counsel's unpredictability and slow response.[4]

  2. 02

    Contracts sit on the revenue critical path. Slow NDAs, MSAs, and DPAs stall sales, onboarding, and partnerships. The fastest, most predictable path to "papered" agreements wins the wedge — and growth-stage SaaS and vendor contracts are the densest, most repetitive flow to start with.[5]

  3. 03

    The law-firm wrapper wins trust as AI scales. Courts have already sanctioned lawyers for AI hallucinations in filings.[2] Buyers demand verifiable, attorney-owned work product. General Legal's attorney-in-the-loop with explicit ownership is the difference between a useful experiment and an enterprise-ready service.

  4. 04

    Three Casetext alumni who built legal AI at scale. Ryan was CTO at Casetext (acquired by Thomson Reuters for $650M in 2023);[3] Javed led AI products for lawyers there; J.P. was a Senior ML and Applied Research Scientist on the same team. Both Javed and J.P. attended Harvard Law and practiced at Fenwick and Cooley. They have already shipped legal AI to enterprise customers — and now they're building the firm.

  5. 05

    Data compounding is the long-term moat. Every reviewed and negotiated contract adds to playbooks (preferred positions), "what's market" by counterparty and segment, and acceptance patterns — improving first-pass accuracy and attorney leverage over time. Incumbent firms don't structure their data this way; tool-only competitors don't sit inside the negotiation loop.[4]

Problem

Outside counsel is slow, expensive, and unpredictable — and contracts sit directly on revenue.

Growth-stage founders face commercial contract review that is slow, expensive, and error-prone. Traditional law firms charge $500–$2,000 per hour without clear estimates, take days or weeks to turn around routine contracts, and often miss harmful terms that can be company-ending — or prevent VCs from counting revenue toward ARR.

This isn't a luxury problem. NDAs, MSAs, DPAs, and order forms gate sales pipelines, customer onboarding, vendor integrations, and partnerships. World Commerce & Contracting research consistently shows organization-wide involvement in contracting and meaningful value leakage from slow, high-touch processes.[5]

The existing fixes don't fix the bottleneck. CLM platforms digitize storage, lifecycle, and workflows — but the negotiation itself, the part where attorney time burns hours per document, remains human-intensive and per-hour-billed.[7][8]

$500–$2,000

Hourly rate at traditional firms

Without clear estimates

Days–weeks

Typical turnaround

On routine commercial contracts

Often missed

Harmful terms

That block ARR or risk the company

Why Now

Three shifts converge: AI gets good enough, accountability becomes mandatory, regulation cracks open.

The window for an AI-native law firm — not just an AI legal tool — opened recently and is closing fast.

AI is good enough to make routine contracting 10× more efficient — but only when wrapped by accountable lawyers.

AI capability. Frontier models can now read a contract, classify clauses, detect counterparty positions, surface the two or three issues that actually matter, and produce first-pass redlines that an experienced attorney can finish in minutes instead of hours. The compute exists; the playbooks exist; the workflows exist. General Legal's attorneys operate at roughly 10× the efficiency of traditional firms using these workflows.

Accountability is mandatory. As AI scales in legal work, buyers want a name on the document. US courts have already sanctioned lawyers for filing AI-generated hallucinations — verification, citation provenance, and human sign-off aren't nice-to-haves, they're table stakes.[2] The firms that win will be the ones that own the work product, not the tools that sit beside it.

Regulation cracked open. Arizona's Alternative Business Structures (ABS) regime permits non-lawyer ownership and economic interests in law firms with court approval — a first in the United States. Other states are running limited pilots, and the regulatory direction is toward opening, not closing. This creates space for venture-backed, AI-augmented "full-stack" legal delivery models that capture proprietary workflow and data advantages.[1]

Align on documents and SLAs, then continuously automate. The economics work because the incentive is to ship the document fast and right — not to bill another hour.
— Sequoia Capital — Training Data podcast, on the per-document law firm model[4]

How It Works

Intake to close: a single loop with attorneys on the critical path.

Intake → first-pass redlines (policy/playbooks) → attorney review and negotiation → counterparty turns → close and knowledge capture.

Step 01

Flat-fee contracting service

$500 per contract covers drafting, review, and negotiation across all turns. Sub-3-hour turnaround SLA. Slack- and email-native interfaces for intake and updates; structured handoff to counsel when needed.

Step 02

AI-augmented workflow

Playbook-driven redlines tuned per customer risk posture. Clause classification and counterparty-position detection prioritize material issues. "What's market" guidance is informed by accumulated counterparty acceptances, industry segment, and company stage.

Step 03

Human-in-the-loop attorneys

Experienced attorneys own outputs, calibrate judgment on the two or three issues that actually matter, and maintain client-specific policies. Escalation paths for edge cases; full audit logs of every change, rationale, and acceptance.

The compounding part: continuous improvement on every contract.

Evals and production monitoring. Accuracy of clause detection, counterparty acceptance rates, turnaround distribution, escalation frequency, and defect rates are tracked across every matter. This mirrors best practices discussed in leading AI engineering forums — and it's what lets General Legal honor the SLA without compromising quality.[4]

"What's market" data compounds. Every accepted and rejected redline, every counterparty position, every closed contract feeds the playbook. Over time, first-pass accuracy improves, the median human minutes per contract drops, and the negotiation guidance the firm delivers gets sharper than anything an in-house lawyer or generalist firm can produce.

Open source as a credibility play. Ryan and the team publish practical tools (e.g., Find The Fuck Up) and a project repository to demonstrate shortcomings in traditional drafting workflows — building public trust and pulling community attention toward the broader thesis.

Market

A $388B services market with the negotiation layer still un-disrupted.

Top-down, US Legal Services value-added was roughly $387.7B in 2024 (nominal) — framing the scale of spend General Legal is addressing.[6] CLM platforms (Ironclad, Icertis) have scaled by digitizing lifecycle, storage, and workflows — but negotiation is the bottleneck General Legal absorbs with guarantees and accountable service.[7][8]

Bottom-up to $10M ARR: at $500 per contract, ~10M ARR requires roughly 20,000 contracts per year (~1,700 per month). That's plausibly supported by 60–120 active customers averaging 15–30 contracts per month, or ~240 customers averaging ~7 per month — consistent with growth-stage SaaS and vendor throughput ranges.

Expansion vectors: more contract templates (SaaS customer/vendor base → procurement and partnering), adjacent jurisdictions and practice add-ons, and deeper data-driven negotiation guidance as the dataset compounds.

Near term — growth-stage tech

Seed to Series C SaaS, vendors, and platforms with repeatable contract throughput. Founder networks and YC W26 distribution drive a referral-led pipeline. Single technical decision-maker; fast decision cycles; high willingness to pay for predictability.

Long term — every contract on the critical path

Expand from SaaS contracts into procurement, partnering, employment, and adjacent practice areas. Move up-market as the dataset and playbook depth justify enterprise pricing. The TAM ceiling is the entire US Legal Services market — $388B — and the negotiation layer is where the rents have always lived.[6]

Contracting work touches the entire organization, and most of the value leakage comes from slow, high-touch processes. ROI from faster cycles is direct.
— WorldCC research summary[5]

Competitive landscape

Six competitor archetypes. General Legal's combination of SLA, flat fee, and attorney ownership beats each on the dimension that matters.

Each archetype has a structural limitation. General Legal's outcome-aligned model is the answer to all six.

Crosby

Direct overlap

Sequoia-backed AI-native law firm; contracts-only; startup ICP. Per-doc model with sub-hour targets and a strong GTM narrative. Head-to-head on ICP and model — General Legal differentiates on SLA reliability, vertical playbooks, and a richer 'what's market' dataset.[4]

Harvey

Tool, not a firm

AI tools for law firms and in-house teams (AmLaw / enterprise). Augments incumbents with research and drafting agents — but no SLA, no flat-fee service, no accountable work product. Indirect substitute, not a competitor on outcomes.[9]

Ironclad (CLM)

Tooling vs. delivery

$150M Series E at $3.2B valuation in 2022. Lifecycle and workflow digitization, contract repository, analytics. But negotiation still burdens the customer — the bottleneck General Legal absorbs.[7]

Icertis (CLM)

Platform-focused

$150M financing in 2022. Enterprise reach and mature CLM stack with AI features. Platform focus, not a per-doc service with guaranteed SLAs.[8]

Cooley · Fenwick · WilmerHale

Incumbent firms

Full-service outside counsel; hourly billing; brand, breadth, and complex matters. Slower turns and cost unpredictability on routine contracts — exactly the gap General Legal targets.

Axiom (ALSPs)

Not AI-native

Alternative legal services provider. Cost-effective staffing and enterprise processes — but no AI-native workflow and no guaranteed per-doc SLA or pricing.[10]

Flat-fee incentives plus AI plus attorney oversight compress turnaround and cost while preserving trust — positioning General Legal as the default commercial contracting desk for growth-stage tech.
— General Legal positioning

Traction

20+ growth-stage logos. Repeat volumes. A referral pipeline through YC W26 and the founder network.

Public: over 20 growth-stage companies have moved their commercial contracting to General Legal.

Positioned publicly as $500 flat fee per contract (all turns) with a sub-3-hour turnaround. Early usage signal: repeat volumes are showing up across initial logos, and customers consistently call out pragmatic attorney guidance alongside speed as the reason they switched.

YC W26 distribution plus the founders' Casetext, Fenwick, and Cooley networks are seeding a referral-led pipeline directly in the growth-stage tech ICP — the densest, most repetitive contract flow available.

The bet: every growth-stage SaaS company that uses General Legal becomes a referral source. Every closed contract feeds the playbook. Every playbook iteration tightens the SLA and the margin.

Founder deep dive

Three Casetext alumni — including the former CTO — who already built legal AI at scale.

Ryan Walker — the technical center of gravity. Math PhD from the University of Kentucky. Spent more than a decade leading engineering for legal AI products — Chief Technology Officer at Casetext (YC13) from 2018 through the Thomson Reuters acquisition in 2023.[3] At Casetext he led engineering on CARA, the recommendation and search system, and document-analysis tooling. Stayed on as VP Technology for CoCounsel at Thomson Reuters, shipping enterprise legal AI with human-in-the-loop, multimodal, multi-agent architectures. Has more experience shipping legal AI to lawyers than almost anyone in the industry.

Javed Qadruddin — engineer first, lawyer second, engineer again. Coding since age nine. Went to Harvard Law because he liked law, too. Practiced corporate law at Fenwick & West for two years advising technology companies, then switched back to engineering in 2013 because — as he puts it — he'd rather actually build things. Got into deep learning in 2014. Spent years at the intersection of law and AI at Casetext leading AI products for lawyers, and at Meta. Public communicator on the limits and uses of legal AI — writes about how AI tools should augment, not replace, experienced attorneys. The bridge between the two worlds the firm needs to unite.

J.P. Mohler — the product brain and the operating attorney. iOS/Android developer who went to Harvard Law and represented startups and VCs at Cooley and WilmerHale for three years. Senior Machine Learning and Applied Research Scientist on the Casetext / Thomson Reuters Innovation team — built prototype legal AI tools with law firms, courts, government agencies, and AI model developers. Editor on the Harvard Journal of Law & Technology, former Google Policy Fellow, and President of the Harvard Law and Technology Society. The rare profile that can ship product, practice law, and translate between researchers and lawyers in the same conversation.

The shared origin — Casetext. All three co-founders are Casetext alumni who built AI for legal workflows together before the $650M exit to Thomson Reuters in 2023.[3] Ryan ran technology, Javed led AI for products for lawyers, and J.P. was a senior ML scientist on the innovation team. They have already done the hard part once — taking a legal AI product from research prototype to enterprise-grade product used by real lawyers on real matters. General Legal is the same playbook, but instead of selling the AI to law firms, they're becoming the law firm.

On the model. The team's bet is that services-as-software — proprietary AI workflows reducing the human minutes per matter, with expert attorneys owning outcomes — is the right structure for the AI era. Hourly billing rewards slow work; flat-fee rewards automation. A law firm wrapper provides the accountability buyers demand. And a compounding "what's market" dataset from accepted and rejected redlines becomes a moat no incumbent firm or tool-only competitor can replicate.[4]

Founders & team

Ryan Walker

Ryan Walker

Co-Founder & CEO

Math PhD, software engineer, and machine learning leader who has spent more than a decade building AI products for lawyers. Chief Technology Officer at Casetext (YC13) — acquired by Thomson Reuters for $650M in 2023 — where he led engineering on CARA, search, and document-analysis tooling. Continued as VP Technology for CoCounsel at Thomson Reuters, shipping human-in-the-loop, multi-agent legal AI to enterprise customers. Now co-founding General Legal to deliver AI-native contract work directly through a law firm.

Javed Qadruddin

Javed Qadruddin

Co-Founder & Partner

Engineer-turned-lawyer-turned-engineer. Coding since age nine; went to Harvard Law because he liked law too. Represented startups at Fenwick & West for two years, then switched back to engineering in 2013 and dove into deep learning in 2014. Led AI for legal products at Casetext (YC13, exited to Thomson Reuters in 2023) and worked on AI at Meta. Has lived at the intersection of law and AI for more than a decade — uniquely qualified to build the workflows that ship contracts in hours instead of days.

J.P. Mohler

J.P. Mohler

Co-Founder, CPO & Managing Partner

iOS/Android developer turned Harvard Law-trained corporate attorney. Represented startups and VCs at Cooley and WilmerHale for three years. Senior Machine Learning and Applied Research Scientist on Casetext / Thomson Reuters' Innovation team, building prototype AI tools with law firms, courts, and government agencies. Editor on the Harvard Journal of Law & Technology, former Google Policy Fellow, and President of the Harvard Law and Technology Society. Now back to building — this time, a more efficient law firm.

Founder–market fit

All three co-founders are Casetext alumni who built AI for legal workflows together before the $650M exit to Thomson Reuters in 2023.[3] Both Javed and J.P. attended Harvard Law and practiced corporate law at top firms (Fenwick, Cooley, WilmerHale) before returning to engineering and product. Ryan led technology for the most successful legal AI startup of the last decade. This is the team that has already shipped legal AI to enterprise customers — now they're building the firm.

Risks & mitigations

Risk

Services margins at $500 / contract with a sub-3-hour SLA are punishing if automation doesn't compound fast enough.

Mitigation

Proprietary AI workflows already make their attorneys ~10× more efficient. Every reviewed contract adds to playbooks and counterparty data, driving median human minutes per contract down over time. Initial ICP (growth-stage SaaS / vendor contracts) is high-throughput and repetitive — exactly where automation compounds fastest.[4]

Risk

AI hallucinations in legal work product create regulatory and liability exposure — courts have already sanctioned lawyers for it.

Mitigation

General Legal is a law firm, not a tool. Every output is reviewed and signed off by an experienced attorney, with citation/provenance practices, evals, production monitoring, and post-mortems on every defect. The attorney-in-the-loop wrapper is the moat — and the reason buyers will trust them over tool-only competitors.[2]

Risk

Regulatory and ownership constraints — most US states still prohibit non-lawyer ownership of law firms, capping how 'full-stack' the model can be.

Mitigation

Arizona's Alternative Business Structures (ABS) regime explicitly permits non-lawyer ownership with court approval; other states run limited pilots and the regulatory trend is opening, not closing. General Legal operates within state bar rules and uses compliant entity structures.[1]

Risk

Direct head-to-head with Crosby (Sequoia-backed) and Harvey on adjacent ground; CLM platforms reduce outsource demand if in-house teams scale.

Mitigation

Crosby is per-document AI-native too — General Legal differentiates on SLA reliability, vertical playbooks, and a richer "what's market" dataset. Harvey is a tool, not a firm — no SLA, no flat fee, no accountable work product. CLMs digitize storage and workflow but negotiation remains human-intensive — exactly the bottleneck General Legal absorbs.[4][7][8]

What we're watching

  • Conversion of the 20+ growth-stage logos into multi-contract retainers and the resulting median contracts-per-customer-per-month curve.
  • Median human minutes per contract as the playbook and 'what's market' dataset compound — the single most important leading indicator for gross margin.
  • Expansion of the contract template surface area beyond SaaS customer and vendor agreements into procurement, partnering, and adjacent jurisdictions.
  • How regulatory openings (Arizona ABS, state pilots) evolve and whether General Legal capitalizes with venture-backed full-stack structures.

References

  1. [1]Reuters — Arizona OKs nonlawyer ownership of law firms (ABS, first US state)
  2. [2]Reuters — U.S. judge sanctions lawyers who cited fake cases generated by ChatGPT
  3. [3]Reuters — Thomson Reuters to buy AI company Casetext for $650 million
  4. [4]Sequoia — Training Data podcast (Crosby episode, per-document law firm model)
  5. [5]World Commerce & Contracting — Research library (contracting benchmarks and value leakage)
  6. [6]FRED (BEA) — Value Added by Industry: Legal Services (NAICS 5411)
  7. [7]TechCrunch — Ironclad raises $150M Series E at $3.2B valuation (2022)
  8. [8]Icertis — Announces $150M financing (2022)
  9. [9]Harvey — Company site (AI for law firms and in-house teams)
  10. [10]Axiom — Company site (alternative legal services provider)