Reading Crosby, Arcline, and the W26 Cohort.
A new category of legal services provider has emerged in the last 18 months. The category does not sell software to law firms. It sells legal services directly to companies, combining AI-assisted drafting with same-day turnaround and fixed-fee pricing.
Two of these providers are now the most-discussed names in the U.S. legal market: Crosby, a registered AI-first law firm out of New York with $85.8 million in venture funding and the law firm Cooley among its investors; and Arcline, a Y Combinator W26 marketplace operating under parent entity Parlai, Inc. Artificial Lawyer terms the category NewMods — new-model AI-first law firms.
This note is a field guide for anyone making a decision in or adjacent to this category. It does not advocate for any side of the conversation. It establishes the verified facts about each provider; maps the stakeholder positions in the market; lays out decision frameworks for each position; and identifies the lifecycle stages where independent advisory adds value.
The analysis assumes the NewMod category will continue to expand into the practice areas served by Boston-area incumbents. If the category instead remains narrowly confined to seed-stage and venture-backed startup contract work — below the engagement thresholds of full-service firms — the urgency we describe is overstated. We address that scenario explicitly in section 5.
The reader who benefits from this note is the partner at a Boston-area law firm asking whether to engage, partner, or replicate; the general counsel asking whether to consolidate outside counsel spend toward a NewMod; the investor asking where the value will accrue in 24 months; the operator at a NewMod planning Boston market entry; and the company architecting a NewMod-arbitrage model with its existing law firm. Each has a real decision to make. None of them has a default answer.
Most discussions about AI in legal services have been about software tools — Harvey, Spellbook, CoCounsel, Hotshot, Lexis+ AI. Those are products that lawyers use inside existing firms. The procurement question is which tool, at what price, with what integrations.
A NewMod is structurally different. NewMods are firms — sometimes registered law firms, sometimes marketplaces — that have rebuilt the legal services delivery model around AI as the primary worker and lawyers as the supervisor. Pricing is per-document, not per-hour. Turnaround is measured in hours, not weeks. The operating model depends on AI doing roughly 80% of the routine work, with attorneys focused on the final 20% of judgment, negotiation, and edge cases.
Y Combinator funded three NewMods in its Winter 2026 cohort: Arcline, General Legal, and LegalOS. Crosby came out of stealth in mid-2025 and has raised three venture rounds since, the most recent a $60M Series B co-led by Lux Capital and Index Ventures. The category is now well-defined enough that Artificial Lawyer, Y Combinator, and the BigLaw investment community are using a shared vocabulary to describe it.
The bear case deserves the same airtime as the bull case. Crosby's $1B in processed contract value represents roughly one-third of one percent of the $300B+ U.S. legal services market. The category's structural significance depends not on its current scale but on its trajectory, on whether the operating model holds at scale, and on whether sophisticated capital continues to underwrite expansion. A reasonable reader can conclude either that the trajectory continues and the category becomes a permanent layer in legal services, or that the trajectory plateaus at the venture-startup commodity band and the broader market remains undisturbed. This brief is structured to remain useful under either outcome.
The strategic point: a NewMod is not a tool an existing law firm buys. It is a different kind of firm, occupying a different position in the market — and depending on which seat the reader occupies, that fact creates an opportunity, a competitive question, a pricing question, an investment question, or a regulatory question. Frequently several at once.
| Field | Value |
|---|---|
| Legal entity | Registered law firm with malpractice insurance |
| Founded | 2024 |
| Headquarters | New York, NY |
| Founders | Ryan Daniels (CEO, ex-Cooley, Stanford JD); John Sarihan (CTO, ex-Ramp tech lead) |
| Total funding | ~$85.8 million |
| Most recent round | $60M Series B (March 2026), co-led by Lux Capital and Index Ventures |
| Investors | Sequoia Capital, Bain Capital Ventures, Index Ventures, Lux Capital, 01 Advisors, Cooley, Elad Gil |
| Disclosed customers | Cursor, Clay, Ramp, Cartesia, UnifyGTM, Rogo |
| Volume processed | $1B+ in negotiated contract value as of March 2026 (vs. $30M at stealth exit in mid-2025) |
| Scope of work | NDAs, MSAs, DPAs, vendor and customer agreements |
| Pricing | Fixed fee per document. Specific rate card not publicly disclosed. |
Crosby's positioning is that the bulk of legal contract work lives in a "dead zone" — too important for a company to skip but not valuable enough for a $1,000-per-hour BigLaw lawyer. Rather than build software to help existing firms close that gap, Crosby built a law firm to close it itself. Its lawyers are employed, not freelance. It carries its own malpractice insurance. Its operating metrics are TTAT (total turnaround time) and HuRT (human review time).
Cooley is on the cap table. Cooley is one of the largest tech-focused BigLaw firms in the country. That investment can be read as a hedge, as a referral pipeline, or as a structural belief that the NewMod model becomes a permanent layer in legal services. The reading depends on the reader's prior. The fact itself is unambiguous: a top-tier BigLaw firm has placed capital alongside Sequoia and Lux in this company.
The volume curve is steep. Cumulative negotiated contract value moved from approximately $30M to over $1B in nine months. That is a 33x increase. Volume curves of this slope reset competitive dynamics in adjacent markets within 12–24 months when they sustain.
The capitalization implies trajectory. $85.8 million in venture funding requires a path to several hundred million dollars in revenue. That trajectory is not achievable on per-document pricing to seed-stage startups alone. Crosby has incentive and capital to expand into mid-market and lower upper-market work over the next 24–36 months. Whether that materializes — and how it lands in Boston specifically — is the empirical question every stakeholder should be tracking.
| Field | Value |
|---|---|
| Legal entity | Marketplace under parent Parlai, Inc. Not a law firm. Lawyers are independent contractors. |
| Founded | 2023 |
| Headquarters | San Francisco |
| Founders | Pamir Ehsas (CEO, ex-outside counsel to OpenAI, Oxford-trained); Stefan Mandaric (CTO, 8+ years ML/data science incl. Omny, MIT Fulbright) |
| YC batch | Winter 2026 |
| Funding | ~$500K disclosed seed |
| Network | 50+ lawyers across U.S., EU, Australia |
| Lawyer pedigree | Cooley, Goodwin, Fenwick, Gunderson Dettmer, A&O Shearman alumni |
| Disclosed customers | 50+ venture-backed startups |
| Compliance | SOC 2 Type 2, ISO 27001, GDPR, SAML SSO |
| Document | Price |
|---|---|
| NDA | $300 |
| Contractor agreement | $500 |
| Offer letter | $500 |
| SAFE review | $500 |
| IP assignment | $900 |
| Partnership / referral agreement | $1,000 |
| MSA | $1,500 |
| Terms of Use & Privacy Policy | $1,500 |
| Intercompany agreement | $1,500 |
| Floor | $200 per document |
Arcline differs from Crosby in three structural ways. Each matters depending on the reader's seat.
Arcline is not a law firm. The disclaimer at the bottom of arcline-ai.com is explicit: Parlai, Inc. and its subsidiaries are not law firms and do not provide legal advice. Legal advice is provided by the independent attorneys clients elect to work with. That structure means liability sits with the individual attorney; privilege and conflict screening become per-attorney problems; malpractice coverage is per-attorney rather than firm-wide. For a corporate buyer, this is something to diligence. For a law firm partner considering referral, this is a different bar-rules analysis than referral to Crosby. For an investor, this is the marketplace-vs.-firm structural debate.
Arcline operates on a different economic model. $500K of disclosed seed capital deployed against an independent-contractor network produces a much lower fixed burn than Crosby's employed-lawyer model. Runway cannot be sized precisely from public data, and the marketplace structure can scale capacity faster than an employed model can. The vendor financial risk is real but harder to quantify; legal AI vendor shakeout is a recurring 2026 industry-analyst theme, though distinct from the NewMod-as-law-firm question this brief addresses.
The price floor sits below traditional firm engagement thresholds. A $300 NDA is below the price at which most full-service firms will open a matter. That price point creates demand that did not previously exist at any incumbent firm — work that companies were either skipping, doing themselves with templates, or sending to LegalZoom-tier providers.
The investor lists across both companies are worth a separate read. Crosby has Sequoia, Index, Lux, Bain Capital Ventures, 01 Advisors, Cooley, and Elad Gil. Arcline has Y Combinator. Neither set is conventional for a legal services provider — and that itself is the signal. Sophisticated venture capital, including a tech-focused BigLaw firm, is treating this category as a software-economics opportunity rather than as a professional-services investment.
Hedge. Smart capital placing small bets across the disruption thesis to maintain optionality.
Pipeline. NewMods as feeders for matters that mature into traditional firm work.
Structural belief. NewMods as a permanent layer in legal services that captures the routine band of the work, with traditional firms moving up-market into judgment and complexity.
The third reading is consistent with how investor capital is allocated when a category is forming. Whether that reading proves correct is the empirical question of the next 24 months. The point for this brief: the consensus view among capital allocators is not "ignore." The right working assumption for any decision-maker is that NewMods become a structural feature of the market — and to plan accordingly.
This is a category that creates decisions for at least five distinct stakeholder types. Each has a different seat and a different question.
Boston-area law firms. Mid-tier and tech-adjacent firms have direct exposure to NewMod competition on routine contract work. General-practice and biotech-pharma firms have indirect exposure as their corporate clients evaluate their own legal stack. Specialty litigation and IP firms sit largely outside the immediate scope but should monitor the trajectory.
General counsel and in-house legal teams. The decision is not "should we use AI" but "where does the work belong." The October 2025 ACC/Everlaw Generative AI's Growing Strategic Value for Corporate Law Departments survey found GenAI adoption among U.S. in-house teams more than doubled in one year, from 23% to 52%, and 64% of respondents expect to rely less on law firms because of internal AI capability. Routing routine work to a NewMod may be cheaper, faster, and more transparent than the existing outside-counsel arrangement. It also requires a new vendor management capability and a new conflict-screening layer.
Companies and operators. Beyond legal teams, every operator — CFOs, heads of business operations, founders — has reason to evaluate NewMods directly. Many are doing so without their incumbent law firms in the loop. The ratio of NewMod-routed work to traditional-firm work inside any growth-stage company is a metric that did not exist 18 months ago.
Investors. Legal services is a $300B+ U.S. market that has historically resisted technology disruption. Capital allocators are now active in the category. The questions are which model wins (employed-firm vs. marketplace), which segments mature first, which incumbents adapt, which become acquisition targets, and which become acquirers.
NewMod operators themselves. Crosby's capitalization implies geographic and segment expansion across the next 24 months, with the Northeast as a logical near-term target given New York domicile. Arcline and the W26 cohort are building distribution. Their decisions about market entry, vertical specialization, partnership structures, and pricing all affect the others on this list.
The right next move is different for each seat. The frameworks below are not exhaustive and are not recommendations — they are decision logics that should produce a coherent answer when populated with the reader's specific facts.
Five postures, not three. Most firms will combine two or three.
| Posture | When it applies |
|---|---|
| Monitor | Specialty or low-exposure firms whose practice mix sits outside the NewMod scope today |
| Defend | Firms with strong relationship-based practices that can compete on judgment, complexity, and depth rather than price and speed |
| Partner | Firms that want to capture upside from NewMods on overflow and commodity work without absorbing the operational build cost |
| Replicate | High-exposure firms with the partner alignment to build NewMod-style capability internally |
| Divest | Firms where commodity contract work is structurally unprofitable and can be routed away cleanly |
Each posture has its own bar-rules and economics analysis. None is right for every firm.
The framework is a four-quadrant work-routing decision: by sensitivity (low/high), by frequency (occasional/recurring), by complexity (routine/judgment), and by client expectation (visible/internal). Most legal departments have a default routing today that no one has rebuilt for the NewMod-available world. Rebuilding it tends to identify both cost savings and quality gaps.
This is the model in which a corporate client's law firm uses a NewMod for first-pass drafting, then the firm's attorney reviews, signs off, and delivers to client. The economics can be attractive for both sides — but the architecture has to clear Massachusetts Rules of Professional Conduct 1.5 (reasonableness of fees), 5.3 (supervision of nonlawyer assistance), 5.4 (fee-splitting with nonlawyers), 5.5 (unauthorized practice), 1.6 (confidentiality of client information sent to a vendor), and the disclosure expectations under 1.4. Practitioners outside Massachusetts should consult local bar counsel — most state rules track the ABA Model Rules but specifics diverge.
Done well, the structure is a margin-expanding service redesign. Done poorly, it is a regulatory and malpractice exposure for the firm and a value question for the client.
The category-level questions: does the employed-firm model (Crosby) or the marketplace model (Arcline) prove more durable? Which adjacent practice areas open next (immigration, employment, trademark, tax compliance)? Where do BigLaw firms invest, partner, or compete? What does insurance and bar-association response look like? Which NewMods reach acquisition by Thomson Reuters, Wolters Kluwer, LexisNexis, or by larger NewMods?
The Boston market has specific characteristics: a heavy biotech and pharma anchor, a meaningful financial services and asset management base, a large higher-education and life sciences research community, and a different VC density than New York or San Francisco. Generic startup-contract positioning translates imperfectly. Vertical depth (biotech licensing, NIH-grant compliance, university tech transfer) and channel partnerships (with VCs, accelerators, research institutions) tend to matter more in Boston than in primary tech-hub markets.
Crosby and Arcline are the most-discussed names but not the only entrants. The Y Combinator W26 cohort included three legal NewMods, two of which are not yet in the Boston conversation but will be within 12 months.
General Legal. Founded by three former Casetext and Thomson Reuters operators: Ryan Walker (ex-CTO of Casetext), Javed Qadrud-Din (ex-Head of AI at Casetext, Harvard Law), and J.P. Mohler (Harvard Law, ex-Cooley and WilmerHale). $11.5M combined seed and pre-seed announced. AI-native law firm focused on commercial contracting for growth-stage companies, $500 flat per contract, sub-hour turnaround via Slack. Earlier stage than Crosby with comparable founder pedigree.
LegalOS. AI-native immigration law firm. Out of scope for most Boston transactional firms but important as a category signal — immigration is a high-volume, rules-driven practice where NewMod economics are particularly favorable. Expect similar models to emerge in employment, trademark, and tax compliance.
Arcline. Profiled in Section 04. We expect the marketplace model to either consolidate into a Crosby-style employed-firm structure or be acquired by a larger legal services platform within 18 months.
The pattern across the cohort is consistent. Founder profiles combine a credentialed practicing lawyer with a senior AI engineer. Pricing is flat-fee per document. Initial scope is contract review and routine corporate work. The expansion path is upward into more complex matters and outward into adjacent practice areas.
Decisions in this category move through five lifecycle stages. The advisory questions at each stage are different, and the cost of getting them wrong compounds across stages.
| Stage | Questions that matter | Independent advisory role |
|---|---|---|
| Pre-decision | What is happening in the category? Who are the credible providers? What are the structural risks? Where does our firm/company/portfolio sit? | Market intelligence, vendor scorecards, exposure diagnostics, scenario modeling |
| Architecture | How do we structure the engagement? What does the contract look like? What is the routing logic? What are the bar-rules and disclosure architectures? | Engagement design, vendor selection criteria, partnership architecture, retained bar counsel coordination |
| Implementation | How do we operationalize this? What is the QA layer? What are the KPIs? Who owns vendor management? | Implementation oversight, operating model design, KPI architecture |
| Governance | Are we delivering quality? Are we capturing the value we projected? Are we exposed where we did not intend to be? | Quality assurance, performance review, vendor scorecard refresh, compliance audit |
| Exit / renegotiation | Has the vendor remained viable? Are we still getting value? Do we need to switch vendors, repatriate the work, or restructure the relationship? | Vendor financial monitoring, alternative provider scoping, transition management |
The five stages exist regardless of which side of the conversation the client occupies. A NewMod entering Boston needs market intelligence and partnership architecture. A law firm structuring a referral relationship needs engagement design and bar-rules architecture. A general counsel evaluating routing needs vendor scorecards and implementation oversight. The frameworks are the same; the client is different.
Three signals will indicate whether the NewMod category is consolidating, accelerating, or stalling.
We will publish updates as these signals develop.
Most decisions fail before the room sits down. The landscape was misread. The contract left room for someone else's interpretation. The intelligence arrived too late to matter.
WLR exists for that moment. It surfaces what matters before it becomes urgent and delivers the integrated view that closes the gap between good decisions and executed ones.
Clients do not call it consulting. They say it is why things went right.
WLR is based in Cambridge, MA. Engaged nationally. The practice spans five integrated service lines: Commercial Strategy & Market Positioning; Business Development & Revenue Execution; Contract & Vendor Strategy; Organizational & Operational Advisory; and Strategic Partnerships & Alliance Development.
WLR does not provide legal advice. Where work touches the Massachusetts Rules of Professional Conduct, qualified bar counsel is engaged as a subcontracted partner.
We arrive before the room.