Employers Save Thousands Per Visa With LegalOS’s AI Platform

From a Family Law Firm to a $1.5M Run Rate in Four Months

LegalOS crossed $1.5 million in annualized revenue inside Y Combinator's Winter 2026 batch, four months after opening for business, putting the AI‑native petition platform in a small club of legal‑tech startups that have moved from zero to seven‑figure run rate inside a single YC cycle. SVPost confirmed the figure, and the company's entry on ycstartups.co corroborated it.

The siblings running it, Matthew Asir and Rachel Asir, grew up inside their father's immigration practice, a 40‑year‑old shop. The third co‑founder, CTO Claire Jutabha, brings a different résumé. The Y Combinator company page notes she shipped LLM safety pipelines at TikTok that flag thousands of high‑risk accounts daily, built document‑classification models at Notarize, and worked at NASA JPL on the Mars Exploration team before joining LegalOS to build its 24 specialized AI agents. The three co‑founders work out of San Francisco.

Matthew Asir had already exited one immigration startup. Founderland's profile records that he previously founded Legal Bullet, a SaaS tool that reduced attorney fees by up to 60%, and landed on the Forbes 30 Under 30 list in 2020 at age 22. He studied astrophysics at the University of Chicago.

A Pioneer Fund post on LinkedIn pegged the firm's run rate at $1.2 million after three months; a month later, LegalOS announced on X that it had hit $1.5 million in ARR inside the W26 batch. Backing came from Y Combinator, Pioneer Fund, and Rice Capital. YC's W26 cohort also included two other AI‑and‑lawyer "NewMod" startups, General Legal and Arcline, which Artificial Lawyer flagged as a deliberate accelerator push into fixed‑fee, AI‑augmented legal services.

LegalOS posted a GTM management consultant role in San Francisco at $140,000–$180,000 to help it move from founders running sales themselves to a repeatable commercial operating model. That hire, more than the revenue number, is the clearest signal that the firm intends to scale beyond its current three‑person team.

The 48‑Hour Filing Engine

LegalOS built its filing engine around a dataset of 12,000 successful immigration petitions, then wrapped that corpus in specialized AI agents that draft narratives, organize evidence, and pre‑flight each case against recent USCIS adjudication patterns. The pipeline produces a complete work‑visa package that the firm says it can deliver in as little as 48 hours, with a licensed attorney (averaging 40‑plus years of practice) reviewing and signing every filing before it reaches USCIS.

Three workflow steps account for the compression. Petition drafting: AI agents generate the narrative and supporting argument, and an attorney edits rather than authors from scratch. Evidence organization: the system structures exhibits and supporting documents into a USCIS‑ready packet. RFE anticipation: LegalOS runs each draft against current USCIS adjudication trends to flag weak documentation before filing. The firm's launch materials describe the result as the difference between "1500 retainers, weeks of back and forth" and a 48‑hour turnaround.

Cost falls out of that compression. Y Combinator's launch post places conventional firms at $10,000–$35,000 per case. The unit math, three employees supporting $1.5M in annualized revenue, implies a per‑case cost well below the traditional band. The gap matters most for the categories LegalOS handles: O‑1, H‑1B, EB‑1A, EB‑2 NIW, L‑1, and TN visas.

The approval outcomes reinforce the speed claim. SVPost noted that LegalOS has filed dozens of applications with a 100% approval rate so far, and the company's own site lists a 99.8% success rate across more than 10,000 visas handled by its attorneys. The firm's O‑1A screener, which produces a preliminary eligibility assessment in about two minutes, extends the same logic upstream.

Competitors are already closing in. A LinkedIn commenter running MeritDraft claimed their tool generates a full O‑1 petition draft in about 15 minutes, complete with AAO precedent citations and 8 C.F.R. references. LegalOS's 48‑hour turnaround is the human‑reviewed wrapper around a machine‑drafted product; rivals are racing to shrink that wrapper further.

What 48 Hours Buys a Rocket Company or a Chipmaker

For a rocket company that just won a Starship flight and a chipmaker racing to ship next‑gen EUV machines, the bottleneck is the green light to put a foreign‑born engineer on the payroll. USCIS reports that the H‑1B regular cap and the 20,000 master's cap for FY2027 have both been reached. Add a $100,000 supplemental payment that briefly applied to new H‑1B petitions between September 21, 2025 and a June 8, 2026 court order vacating the agency guidance. With DHS signaling it will resume collection if the order is lifted, every extra week an immigration filing spends sitting on a lawyer's desk is a week a critical hire can't start.

The visa mix LegalOS markets — H‑1B specialty‑occupation petitions, O‑1 "extraordinary ability" cases for founders and senior engineers, EB‑2 NIW self‑petitions, and EB‑1 green‑card pathways — maps onto frontier‑tech hiring profiles. Wisa's analysis and PriorityPath's employer guide show SpaceX sponsors roughly 10,000–15,000 H‑1B filings a year across Hawthorne, Starbase, and its other engineering sites, focusing on aerospace, propulsion, avionics, and software roles; prevailing wages for its sponsored aerospace engineers run $140,000–$165,000. The company also backs EB‑2 and EB‑1 green cards for qualifying staff. ITAR review applies to certain export‑controlled roles, and a new Form I‑129 became mandatory in April 2026.

ASML, the Dutch lithography company that anchors the global semiconductor supply chain, posted 47 new openings in the past week on the Zero G Talent board, including a Principal Opto‑Mechanical Engineer and a Senior mixed‑signal electrical engineer in San Jose. Stripe added 61 roles in the same window, including Machine Learning Engineers in South San Francisco at $212,000–$318,000. For these employers, a 48‑hour‑turnaround petition shop isn't a luxury; it's a hiring‑velocity tool.

The harder constraint is the O‑1 and EB route itself. Get Wisa's reporting cites a prevailing‑wage figure of $140,000 for an aerospace engineer at SpaceX in Hawthorne, with another filing for a Senior Propulsion Engineer at $165,000. Those are the documentation stakes that LegalOS's platform is built around.

How Traditional Firms Are Fighting Back

Traditional immigration practices are not standing still while AI‑native entrants like LegalOS compress visa petition turnaround to 48 hours. They are buying the same tools, and quietly rebuilding their economics around them. The 8am 2026 Legal Industry Report, summarized by DocketWise, found that 82% of immigration law professionals personally use general‑purpose AI platforms such as ChatGPT or Claude for work tasks, against 69% adoption across legal practice as a whole. Forty percent of immigration practitioners now use generative AI tools every day.

Individual usage is racing ahead of firm‑level governance. Only 8% of immigration firms operate under a formal, written AI policy; 64% have no policy at all.

The pressure starts with unit economics. DocketWise's survey of 110 firms using AI for immigration casework reported these deployment patterns:

The translation figure matters when over 70% of immigration law clients are Spanish‑speaking.

The reported gains are large enough to reset a firm's headcount plan. A LegalBridge case study documents a U.S. immigration firm that scaled from 150 to 450 active cases with zero new hires. Paralegals moved from 50 cases each to 150, while case completion time fell from four months to six weeks, a 67% reduction. Form preparation collapsed from 3–4 hours per case to roughly 45 minutes of review. Document drafting dropped from 6 hours of paralegal time to 90 minutes of refinement on an AI‑generated first pass. USImmigration.ai's case study reports similar shapes: a 70% reduction in onboarding time, 60% savings on manual document prep, three times the monthly application throughput, and near elimination of missed deadlines.

What is driving the scramble is the math of legacy billing. Leads Under Control estimates attorneys spend roughly 60% of their time on administrative work rather than practicing law. If the average immigration case generates $3,000 to $8,000 in fees, paying attorney‑rate salaries for status‑call reception work is no longer defensible. The legal industry's average follow‑up time sits at 47 hours. AI tools that bring that response window below 6 hours, and cost $299 a month, change the calculation.

Regulators are catching up just as quickly. The Florida Bar's Opinion 24‑1 sets guardrails on AI use in client representation; the Court of King's Bench in Manitoba now requires disclosure of AI use in submissions. With 49% of immigration attorneys citing data security and 40% each flagging trust in outputs and unauthorized‑practice‑of‑law concerns, the firms that survive the next phase will be the ones that wrap AI in documented compliance.

Boundless, Remote, and the New Procurement Layer

LegalOS is not entering a vacuum. Its AI‑native go‑to‑market lands inside an immigration‑services sector that has been consolidating for years, with Seattle‑based Boundless emerging as the most aggressive consolidator. Founded in 2017, Boundless built an online platform that has helped thousands of individuals apply for green cards and U.S. citizenship while providing access to independent immigration attorneys.

The acquisition arc tells the story. In September 2020, Boundless acquired RapidVisa alongside a $7.5 million Series A‑1 led by Foundry Group, bringing total equity funding to over $18 million. CEO Xiao Wang said the move made Boundless "the largest international online immigration company" at the time, and the deal doubled team size and tripled the number of immigration categories supported through technology. A $25 million Series B followed in April 2021. In February 2025, Boundless added Bridge, a U.S. business‑immigration company. Then, in October 2025, Boundless announced its largest move: the acquisition of Localyze, a Berlin‑headquartered employer‑focused global‑mobility platform founded eight years earlier that covers visa and work‑permit processing, residence permits, relocation, business‑trip compliance, and destination services across Europe.

The Localyze deal changes the geometry for LegalOS. Boundless's announcement frames the acquisition as creating "a unified solution" for multinational companies to manage employee visas, relocation, and compliance across the Americas, Europe, and the Asia‑Pacific region. Wang framed the timing bluntly: "Global companies face a maze of immigration uncertainty, and the cost of a delayed or denied petition is immense." Localyze CEO Hanna Asmussen said joining Boundless was "the fastest way to achieve" her team's mission "on a global scale."

Underneath the strategic language, the deal carries real friction. GeekWire reported that Boundless carried out a reduction in force on September 25, 2025, six days before the Localyze announcement, leaving the company with roughly 500 employees; the number of layoffs and their connection to the deal were not disclosed. The companies also did not publicly detail purchase price, financing structure, employee retention, or technical integration.

The HR‑platform partnerships sit on a different layer of the stack but pull in the same direction. Remote.com and Deel have built employer‑of‑record footprints and have been layering immigration and mobility services onto their platforms. None of these HR platforms practice immigration law the way LegalOS or a traditional firm does, but they are increasingly the procurement layer through which frontier‑tech employers buy global‑hiring services, which makes them either channel partners or competitive gateways for AI‑native firms trying to reach the same buyer.

Boundless has telegraphed where it intends to compete next. In its post‑acquisition vision statement, the company said it would use a "proprietary immigration dataset and continuously learning AI" to "route talent to the best pathways across regions, predict outcomes, and benchmark performance," turning immigration "from a reactive necessity to a proactive, strategic advantage."

USCIS's Own AI Reshapes the Filing Game

The federal government is no longer a passive bystander in the immigration AI race. U.S. Citizenship and Immigration Services now runs AI systems tied directly to its adjudication workflow, per the DHS AI Use Case Inventory. The American Immigration Council places the broader DHS deployment at 105 active AI use cases across components. Among USCIS's cataloged tools, the ELIS Evidence Classifier has been linked to a doubling of the share of applications processed within 30 days.

The policy backdrop moved quickly. On August 19, 2025, USCIS published Policy Alert guidance clarifying discretionary factors in certain immigration benefit requests; on November 27, 2025, a second Policy Alert addressed the impact of INA 212(f) on discretionary benefits adjudication. Then came USCIS Policy Memorandum guidance on adjustment‑of‑status and discretion, grounded in Presidential Proclamations 10949 (June 4, 2025) and 10998 (December 16, 2025). Within weeks, on June 5, 2026, the U.S. District Court for the District of Rhode Island vacated three related policy documents in Dorcas International Institute of Rhode Island, et al. v. USCIS, entering final judgment on June 11, 2026. USCIS publicly disagreed with the ruling but said it would comply pending possible further review.

That whiplash is itself a driver of demand for AI‑native immigration services, because lawyers who can ingest new guidance into software faster than a human can re‑read a memo gain a structural edge. Lawcore.ai reports a 67% year‑over‑year surge in immigration law queries on AI search platforms through 2026.

The operational pressure is concrete. SG Legal Group has noted the institutional pressure to clear an 11‑million‑case backlog will keep automation expanding regardless of who's in office. The downstream effect on filings is already visible: firms report same‑day Requests for Evidence on I‑485, I‑130, I‑864, and I‑765 forms that historically arrived weeks after submission, an artifact of automated intake screens flagging inconsistencies within hours of receipt. Cozen has warned that the rapid expansion of AI in processing is driving a surge in RFEs and denials tied to evidence misclassification, data‑matching errors, and heightened social media screening. Zaveri Immigration Law put a cost on it: new screening measures can add months to onboarding for international hires.

The litigation track is open, too. Plaintiffs including labor unions and immigrant‑rights NGOs have asked whether algorithmic flagging is reviewable under the Administrative Procedure Act, whether model logic can be compelled through FOIA, and what remedies exist for bias. USCIS's stated position is that it does not publicly assert AI approves or denies benefits, a posture that leaves the technology deployed but the accountability framework unfinished.

What LegalOS Doesn't Do

LegalOS draws a tight perimeter around employment‑based immigration, and that focus defines what it is not. The Y Combinator company page lists exactly eight petition types it handles: O‑1A and O‑1B extraordinary‑ability visas, EB‑1A, EB‑1C, EB‑2 NIW, H‑1B, L‑1A, L‑1B, and TN. Outside that menu, the company is silent, by design, not by oversight.

Three categories sit clearly outside the firm's scope. Asylum work, which centers on Form I‑589 filings, operates under a different statutory regime entirely: USCIS's Form I‑589 instructions impose a one‑year filing bar after arrival, require the Annual Asylum Fee under Public Law 119‑21 for principal applicants, and prohibit filing while in active Executive Office for Immigration Review proceedings unless the applicant qualifies as an unaccompanied minor. Refugee resettlement, family‑based petitions, naturalization, and removal defense before the immigration courts are similarly absent from LegalOS's offering. The firm's contact page frames its hours as a service window for employment‑based petitions, staffed 8 a.m. to 5 p.m. EST, Monday through Friday — a schedule that signals a transactional, employer‑facing practice rather than an emergency‑response or litigation shop.

The pricing model reinforces the boundary. LegalOS's pitch — built around the Asir family's 40 years of immigration law and 10,000+ visas filed — attacks the $10,000–$35,000 per‑petition fees charged by traditional firms for employment visas. That economic target has no analog in asylum practice, where applicants typically arrive with limited financial resources and where, as Access61's 2025 analysis notes, asylum law demands "specialized knowledge of international human rights law, country‑specific political conditions, and intricate legal and procedural interpretations" that employment‑visa automation cannot replicate.

Litigation is the third exclusion. Every petition LegalOS files is reviewed and signed by a licensed immigration attorney, and its scope ends at USCIS adjudication, not at the immigration court bench. For frontier‑tech employers weighing LegalOS against a full‑service firm, the practical implication is that any matter touching humanitarian relief, family‑based adjustment, citizenship, or contested hearings still requires a separate practitioner.

USCIS's AI Demand Keeps Rising

USCIS's own automation is rewriting the incentives that once protected legacy immigration practices, and LegalOS is one of the first AI‑native firms built from scratch for that environment — one whose filing engine, trained on 12,000 successful petitions, can answer USCIS's ELIS Evidence Classifier in the same algorithmic register that screens it. The question for the next twelve months is no longer whether AI belongs in immigration law but whether the firms that adopt it fastest can keep their judgment intact as the machines on both sides of the filing get smarter together.

Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.