Norm Law: The Client Layer That Stays When Lawyers Leave

Key points

• A client layer built from past deals keeps client knowledge in place when lawyers leave.
• Changes to an agent get a peer review from another legal engineer, like code changes in software.
• Each matter is priced by outcome in advance, so the firm takes only work whose scope can be fixed.

Two linked organizations aimed at the legal work of financial institutions

In a typical big law firm, when work on a deal or a regulatory review begins, associates are put on it. Associates are the junior and mid-level lawyers at a firm. They read huge piles of contracts and write the first draft. Then an experienced partner reviews the result and deals with the client. Under billing that follows hours worked, the working hours of the junior layer that wrote drafts have long been a steady pillar of big-firm profit. But what happens to the way a firm works if software agents, not people, handle the drafting stage, and lawyers focus only on building and supervising those agents?

Norm Ai is a legal and compliance AI company that John Nay founded in New York in mid-2023. Compliance means the work of making sure a company strictly follows laws and internal rules. The company started by selling AI agents to the in-house legal and compliance departments of banks, asset managers, insurers, and hedge funds. For example, Blackstone, a large private equity firm, used Norm Ai in its in-house legal and compliance teams, mainly to review content subject to financial regulation.

On November 20, 2025, Norm Ai announced the launch of Norm Law LLP, which it calls an AI-native law firm. Starting with financial services, Norm Law offers institutional clients the work a traditional corporate law firm does. According to Nay on a July 2026 podcast, its current work covers real estate transactions and finance, private funds, registered funds, private equity, venture capital, and matters involving the US Securities and Exchange Commission (SEC). Norm Law has actively recruited senior partners from big law firms. Mike Schmidtberger, who chaired the executive committee of the big firm Sidley Austin from 2018 to 2025, has joined as Norm Law’s chairman.

The two organizations are legally separate but closely tied. Norm Ai is a software technology company that builds the agents. The lawyers at Norm Law, a law practice, do the actual legal work as outside counsel representing clients. According to a Norm Ai job posting, Norm Law is “an independent, AI-native law firm and is the only law firm authorized” to use the Norm Ai platform in production legal work.

Former lawyers build the agents, and senior lawyers give the final approval

At the center of this way of working is a job called the legal engineer. Legal engineers are former lawyers who stopped practicing law and now build agents directly. In an October 2025 article, Norm Ai said it had more than 30 legal engineers. A month later, in the press release for Norm Law’s launch, it said more than 35. All are former lawyers. On the podcast, Nay described the mix: “most of them are around the fifth-year associate level from other law firms.” He added that “they basically become software engineers” and “They’re not practicing law.”

The legal engineer’s role has moved closer and closer to software code over time. In October 2025, Norm Ai introduced its own tool, the Legal Engineering Automation Platform (LEAP), as a no-code platform that lets lawyers build agents without a software engineer. But a later job posting, while not requiring coding experience, said that after a few weeks of training, “Many of our Legal Engineers are writing code in the terminal.” The requirements were a law degree from a top US law school, at least three years of legal practice, and hands-on AI fluency. The base salary ranged from $175,000 to $225,000.

The work of a legal engineer in the posting is clear. They turn complex statutes, administrative rules, enforcement cases by authorities, and legal doctrine into structured, testable systems that an agent can understand and run. They build agent prototypes on the platform and find missing features through real client workflows. Then they hand reusable core pieces to software and AI engineers to build as proper features. They also test agents on real client data with former regulators and senior practitioners, and sometimes sit on site at the client’s office to help.

When a new matter comes in, the work is handled in layers. For each type of matter, such as a deal review, a base agent is prepared in advance, built from the knowledge of senior partners. Partners sit down face to face with legal engineers, explain their practical workflows in detail, and put their experience and expertise into the base agent. On top of that goes a client layer, trained on the individual client’s past deals, bundles of contracts, and internal precedents. This client layer gets thicker the more of that client’s work it handles. Nay calls the result a digital associate and explains, “now this digital associate is there, and it’s not going to leave.”

When a real deal starts, agents, not people, handle the documents first. According to Nay, in one deal the documents were “analyzed by an agent behind the scenes before a human even stepped in at all.” Norm Law’s own description, quoted by Bloomberg Law, is: “AI agents developed by attorneys complete initial drafts of work, and lawyers supervise and review the outputs.” A press release on August 11, 2026 describes this as “senior attorneys supervising, calibrating, and improving the AI agents.”

How the two organizations handle work

The checking system for work in Norm Ai’s materials has five layers. The first is peer review among legal engineers. When a change alters how an agent interprets and applies a legal policy, it must be reviewed in advance by another certified legal engineer before it is used in real work, just as a software team does peer review before it ships code.

The second is testing with experts. Legal engineers verify agents on real client work together with former regulators and senior practitioners. Norm Ai’s advisory group includes a former SEC commissioner and a former head of New York State’s financial regulator.

The third is explainability. Norm Ai says it provides an explanation of the reasons for every determination an agent makes, so that people can check why. The fourth is a supervisor agent that watches agents. Norm Ai has built supervisor agents that monitor how other AI agents work in regulated settings. However, the materials describe this supervisor agent mainly as a tool for watching client companies’ AI systems, and it is not confirmed whether it directly reviews Norm Law’s legal drafts.

The fifth is the final approval of a licensed lawyer. Nay said, “nothing goes out the door without that human supervision by a barred attorney.”

The way it gets paid is completely different too. Norm Law does not use hourly billing, which charges by time worked. Nay said, “We primarily price by outcome.” It prices by unit of work, such as one deal or a bundle of contracts to review. It uses what the client spent in the past on similar legal work as the basis and says, “We align on a price ahead of time.” Nay admitted that setting the right price in advance is tricky, and that it “does limit the focus, at least initially, of the types of matters that we’re going to do.” So both sides take only matters where the scope of work is clearly understood, and the firm seeks long-term relationships with large clients so that differences in workload between individual matters average out. Nay said the company is “not focused on bet-the-company AI agents for bet-the-company work.”

Keeping in software the precedent that used to live in people’s heads

The first point where Norm Law splits most from an ordinary corporate law firm is the role of junior staff. At a traditional firm, junior and mid-level associates take on most of the basic work, such as document review, drafting, and checking against past precedent. At Norm Law, agents do that first pass. A lawyer of about fifth-year level, who would have been writing drafts at an ordinary firm, works at Norm Ai as a legal engineer who designs and maintains the agents. Partners keep client advice and review of the output, and they spend time teaching their professional judgment to the agents. Norm Law has also brought in partners and some senior associates. But according to the company’s account reported by LawSites, senior lawyers supervise agents instead of managing the traditional associate pyramid. In other words, the labor of the junior layer that wrote drafts has been replaced by agents.

The second difference is the form in which knowledge builds up in the organization. At an older firm, the details of a client company and its past deal experience mostly stay in the heads of the lawyers who worked on those matters. When such a lawyer moves to another firm, part of the accumulated knowledge walks out the door too. Norm Law, by contrast, stores the client’s own knowledge as software in the agent’s client layer. Even if the person in charge changes, it stays in the system. Justin Rattigan, a partner at Norm Law, said in a May 2026 press release, “The scarce resource in sophisticated legal work is no longer document production.”

The third difference is the organizational split between the technology company and the law practice. The agent software is developed by Norm Ai, a pure technology company whose engineers do not practice law. The actual legal advice and the final sign-off are the responsibility of Norm Law, the firm whose lawyers hold law licenses. As LawSites pointed out, existing legal tech companies such as Harvey and Legora sell software tools that lawyers can use. Norm Ai, by contrast, has its own team of legal engineers, filled with former lawyers, build the agents directly. It delivers the legal service itself to clients through the connected dedicated firm, Norm Law.

The fourth difference is the fee structure and the source of profit. Under hourly billing, pay follows the time spent on the work, so if the work takes less time, the amount billed falls too. Norm Law sets a price for the outcome in advance, using the client’s own past spending as the benchmark. In principle, if agents make the work faster, the firm keeps the shortened share as profit instead of cutting the hours it bills. But the price of this approach is that the work it takes narrows to matters whose scope can be set and priced in advance.

The pyramid shakes, and a new kind of expert job appears

Once a system takes hold in which agents do the drafting and senior lawyers review, the associate pyramid, the basic structure of the big law firm, could lose its economic foundation. Hourly billing has made money on the huge number of hours junior lawyers put in. In an environment that prices by outcome, a firm does not need large numbers of junior lawyers to write drafts by hand. In an interview with Bloomberg Law, Schmidtberger predicted that Norm Law would “coexist very comfortably with Big Law.” That may be true as long as Norm Law handles only routine work with a clear, standardized scope. But if the model gradually widens into more complex deals and a variety of regulatory work, the big-firm business model, which has relied on junior hours, could come under heavy pressure.

At the same time, a completely new career path could open for lawyers. Hiring mid-level lawyers and turning them into engineers who design agents themselves shows that a new job between lawyer and software engineer can exist, one that existing law schools and traditional firms did not teach. A legal organization that wants to build its own agents, instead of simply buying outside tools, will need to secure its own legal engineers, people who understand code and can turn legal doctrine into systems.

Other professional services industries that work with expert knowledge can learn a few operating principles here.

First, keep the experts next to the builders. Norm Ai’s agents come out of the process in which partners go through practical workflows with legal engineers. Having a domain expert who can build agents, or who works every day with the people who build them, may matter more than the choice of tool.

Second, treat changes to an agent like code changes. Peer review before a change is used in real work, testing on real client work, and an explanation of the grounds for every result are an example of a software team’s habits moving into expert work.

Third, set performance-based prices only for work whose scope can be defined. Outcome-based pricing works when both the client and the provider can clearly define the unit of work. Norm Law started with matters whose scope is clear, and Nay said this choice limits the range of work it takes on.

Fourth, keep a named expert who takes responsibility. However the agents are built and tested, the final approval belongs to a licensed lawyer. In regulated work, this boundary is likely to stay in place.

Large-scale expansion and independence are not yet fully verified

Most of what is known so far about the workflow, the agent verification process, and the pricing structure rests on announcements by Norm Ai, Norm Law, and founder Nay. Norm Law’s total number of lawyers, its actual number of clients, and how much people had to correct the drafts the agents wrote have not been made public in a transparent way.

The company’s own materials also give slightly different numbers. The number of legal engineers is given as more than 30 in one place and more than 35 in another, and the size of client assets under management swings between $25 trillion and $35 trillion. All such specific numbers are the company’s own claims. The main clients named publicly, Blackstone, Coatue, and Bain Capital Ventures, are also investors in Norm Ai.

The governance structure between the two companies and how fees are divided are not spelled out in the published primary sources either. According to an analysis by the research firm TwinLadder, Norm Law appears to be owned by lawyers under New York rules and to pay license fees to Norm Ai, the technology company. The analysis points out that the license payments must not turn into prohibited fee-sharing with a non-lawyer entity. It also notes that lawyer review is the main safeguard described, and that the review burden grows as agents write more drafts.

Norm Law says it avoids high-risk and novel matters for now. Whether this approach works outside the defined scope of financial services work has not yet been proven.

---

Sources

1. Norm Ai — Legal Engineering: A Paradigm Shift in Law — https://www.norm.ai/resources/legal-engineering-a-paradigm-shift-in-law — 2025-10-24
2. Norm Ai — Norm Law (open letter by John Nay) — https://www.norm.ai/resources/norm-law — 2026-01-06
3. PR Newswire — Norm Ai Announces $50 Million Blackstone Investment, Launch of New AI-native Law Firm Norm Law — https://www.prnewswire.com/news-releases/norm-ai-announces-50-million-blackstone-investment-launch-of-new-ai-native-law-firm-norm-law-302621622.html — 2025-11-20
4. Built In NYC — Norm Ai, Legal Engineer job posting (removed 2026-03-05) — https://www.builtinnyc.com/job/legal-engineer/7045418 — accessed 2026-10-22
5. PR Newswire — Norm Law Launches Emerging Companies and Venture Capital Practice with Senior Additions from Pillsbury and Bain Capital Ventures — https://www.prnewswire.com/news-releases/norm-law-launches-emerging-companies-and-venture-capital-practice-with-senior-additions-from-pillsbury-and-bain-capital-ventures-302782794.html — 2026-05-27
6. PR Newswire — Norm Ai and Norm Law Open Washington, D.C. Office as Norm Law Welcomes Partners Katelyn Kimber and Lindsey Savage — https://www.prnewswire.com/news-releases/norm-ai-and-norm-law-open-washington-dc-office-as-norm-law-welcomes-partners-katelyn-kimber-and-lindsey-savage-302847662.html — 2026-08-11
7. AI and the Future of Law (podcast) — What is an AI-Native Law Firm? John Nay on Norm AI and Norm Law — https://www.adr.org/podcasts/ai-and-the-future-of-law/what-is-an-ai-native-law-firm/ — 2026-07-21
8. Bloomberg Law — Former Sidley Leader Embraces ‘Future,’ Joining AI Law Firm — https://news.bloomberglaw.com/securities-law/former-sidley-leader-embraces-future-joining-ai-law-firm — 2026-01-22
9. LawSites — Norm Ai Raises $50 Million from Blackstone, Launches AI-Native Law Firm — https://www.lawnext.com/2025/11/norm-ai-raises-50-million-from-blackstone-launches-ai-native-law-firm.html — 2025-11-26
10. LawSites — Norm Ai Hits Unicorn Status with $120M Series C at $1.2 Billion Valuation — https://lawnext.com/2026/07/norm-ai-hits-unicorn-status-with-120m-series-c-at-1-2-billion-valuation.html — 2026-07-07
11. SiliconANGLE — Norm Ai nabs $120M at $1.2B valuation to bring AI agents to law — https://siliconangle.com/2026/07/07/norm-ai-nabs-120m-1-2b-valuation-bring-ai-agents-law/ — 2026-07-07
12. TwinLadder — Norm Law: Blackstone’s AI-Native Law Firm Play — https://www.twinladder.ai/en/blog/norm-law-analysis — 2026-10-22