The MSO Is Turning Law Firms Into Assets. AI Is Setting the Price

For a century, the most valuable asset in a law firm walked out the door every time a partner retired.

That is starting to change. Outside money is buying into law firms for the first time in modern history, and it is paying prices that only make sense if the firm can be rebuilt around software. This has received a fraction of the attention given to the chatbots on display at legal technology conferences, and it matters a great deal more.

In April, Manifest raised $60 million at a $750 million valuation, the largest Series A in legal technology history, to build law firms rather than sell software to them. NormAi followed in July with $120 million at a $1.2 billion valuation, an AI-native law firm attached, and Blackstone Inc. as both investor and customer. Then in August the Financial Times reported that Charlesbank Capital Partners, the $24 billion fund that acquired accounting firm Aprio in 2024, was in advanced talks to take a stake in Wood Smith Henning & Berman, a 550-lawyer insurance defense firm, at an enterprise value of about $700 million against roughly $46 million in projected 2026 earnings.

That is about 15 times earnings. Until recently, the comparable figure for a law partnership was zero. Retiring partners sell nothing. They collect back the capital account they paid in years earlier, returned without appreciation. In the entire history of the professional partnership, the exit multiple on accumulated expertise has been zero.

The Structure Behind the Deals

The vehicle enabling these transactions is the management services organization, or MSO. Rule 5.4 of the American Bar Association's Model Rules bars non-lawyers from holding an interest in a law practice or sharing in its fees. The MSO leaves that rule in place and works within it.

Under the structure, the law firm remains lawyer-owned and keeps what the rules require it to keep: client relationships, legal judgment, privilege and sign-off on work product. A separate services entity holds everything else, including technology, back-office operations and, increasingly, the encoded workflows that do the work associates used to do. The firm pays the entity a fee for those services, set by independent valuation rather than as a percentage of legal revenue. Outside investors buy into the services entity.

Medicine adopted the same structure a generation ago under an identical prohibition on non-physician ownership, and roughly a third of US dental practices now operate under one. Law is adopting it much faster. Holland & Knight LLP's legal services transactions team has closed more than 30 law firm MSO deals across 20 states this year and has roughly 100 more in progress, according to Law.com. The queue is long enough that new firms wait weeks for a call back.

Nor is the interest confined to mid-market firms. The Financial Times reported in August that Paul, Weiss, Rifkind, Wharton & Garrison LLP, Quinn Emanuel Urquhart & Sullivan LLP and Proskauer Rose LLP have each held talks with private equity groups or their advisers about structures permitting outside investment. None has denied the conversations. John Quinn, Quinn Emanuel's founder, told Bloomberg Law that capitalizing his firm's $2.5 billion to $3 billion in revenue at even a conservative multiple would make it "worth $10 billion-plus as an enterprise," while conceding that persuading partners is "a generational conversation."

Why Capital Alone Failed Before

Investors have had access to legal software for a decade, and the results were underwhelming. Copilots sold by the seat made lawyers faster at work that was billed by the hour, which meant every gain in efficiency shrank the invoice. A firm that halves the hours a matter takes has halved its revenue on that matter.

Capital on its own buys a more expensive version of the old pyramid. An investor who purchases a share of back-office cash flows and installs better billing software owns a back office in a wrapper, and its multiple is capped by the same arithmetic that has always capped services businesses. AI on its own runs into a different wall. Encoding a firm's precedent into production-grade infrastructure requires engineering investment, and a partnership that distributes all of its profits every year and carries no retained earnings has no way to fund it. The legal profession has never had an R&D line.

Where the two combine, something structurally different appears. Outside capital pays for engineers to work inside the firm. Their work drives the marginal cost of a matter down as volume grows, in the way manufacturing costs fall along a learning curve. A firm that knows its cost per matter can quote a fixed price at intake, and fixed prices bring in clients whom hourly rates had kept out of the market entirely. Each new matter then adds to a precedent base the firm owns rather than stores.

Call it the Legal Scaling Loop. It yields a prediction specific enough to be wrong: firms that complete it will grow legal capacity several times faster than lawyer headcount, and a gap of 20% would not count as confirmation.

Early Evidence

The data points are thin but they exist. Hicksons, a Sydney commercial firm in insurance defense, grew from roughly 100 lawyers to 400 in 18 months on an AI-driven cost structure. A chronology that once consumed 60 hours of junior time now takes four to six. The firm quotes flat fees at about 70 cents on the dollar against competitors who can only bill hourly, and firm-wide margins rose by at least 10 points.

Kirkland & Ellis LLP, the highest-grossing firm in the US, is reportedly spending about $500 million to build a proprietary AI platform with Palantir Technologies Inc. That figure is what running this strategy out of partner distributions actually costs, and it explains why only a firm of Kirkland's scale can attempt it without outside capital. Everyone else needs the structure.

On the demand side, a single lawyer running on frontier models recently delivered an appellate brief for about $20,000 that would have cost $350,000 from the firm where that lawyer trained. In one dispute run on August's platform, a law firm completed a matter for under $500,000 against an opposing quote north of $4 million. Complex work turned out to be more tractable than the profession likes to admit. A lawsuit is unique, but its chronology follows a pattern, and the same holds for a company and its financing process.

The Case Against

Three arguments would undermine this thesis, and readers should weigh them.

The strongest is that the surplus from cheaper legal work flows to clients rather than to firms. If every firm can encode its workflows, competition pushes prices toward the new marginal cost and buyers keep the gain, as consumers did when textile prices collapsed in the Industrial Revolution. The honest version of the claim is narrower than the one I would like to make. There is a window, measured in years, in which a converted firm earns software margins behind a labor price because most competitors have not converted. What persists after the window closes is scale, client relationships and the pricing and verification infrastructure that makes fixed-price delivery underwritable. Anyone buying this sector expecting permanent 60% delivery margins has mispriced it.

A second objection is regulatory. In 48 states, the MSO remains a workaround to a rule that has not been repealed. State bars, the plaintiffs' bar and incumbent firms all have incentives to challenge it, and a single adverse ruling in a large state, or one malpractice scandal at a private equity-backed firm with an AI-attributable error, could change the regulatory temperature quickly. Everyone underwriting these deals is underwriting that risk, priced or not.

A third is that this is simply a private equity roll-up wearing an AI costume. Charlesbank's sequence, from physicians to dentists to veterinarians to accountants and now lawyers, is a well-understood trade requiring no language models. The test is revenue per lawyer over the next three years. Flat revenue per lawyer alongside growth by acquisition would mean the roll-up reading is correct. A move from roughly $750,000 toward $1.5 million or more would mean something else is happening.

What Firms Should Do

The decision in front of firm leaders is whether to change the capital structure. Which copilot to license is a smaller question, and it comes second, because the engineering cannot be funded from partner draws. Any prospective sponsor should be asked two things: what is the plan for revenue per lawyer, and do the lawyers hold equity in the entity being bought? If the answers are "consolidation" and "no," the firm is selling into a back-office roll-up and should price it accordingly.

For general counsel, the recommendation is more uncomfortable. Run the last three significant work products through a frontier model this week. What you find will tell you more about the state of this market than this article can.

A few dated predictions, so this can be checked. At least one AmLaw 50 firm will publicly announce an outside-capital structure by the end of 2027. Law firm MSO and alternative business structure formations will exceed 1,000 across US jurisdictions by the end of 2028. Incoming associate classes across the AmLaw 100 will be down at least 25% from 2025 levels by the end of 2029.

Whether those land or not, the underlying claim rests on economics every partner already knows from the inside. A partnership has never been able to own the expertise it builds, and the MSO, with encoded workflows inside it, is the first structure that lets it. The firms working out how to do this are not waiting for the story to become public.

Thomas Bueler-Faudree is the founder of August, a legal AI company that works with law firms on management services structures and holds equity in several of the entities described in this article.