You Have Until 2027 To Choose Your Firm's Form

LLP vs ABS and the Impact of Retained Earnings on Growth

The most-discussed legal-tech story of the year is the wrong story.

The shift that matters is not Claude-for-Legal. It is not Harvey at $11 billion. It is not the 25,000 attorneys who attended a single Anthropic webinar earlier this year, although that figure happens to equal the entire global membership of the International Legal Technology Association. The shift that matters is the one nobody on that webinar mentioned.

Legal services are moving from a labour business to a capital business, and the partnership form cannot finance a capital business.

This is not a technology argument. It is an accounting argument that AI has made unavoidable. I came to it the long way round, having spent the past few months in conversations with managing partners, in-house counsel, and a steady drip of three-to-five-year associates considering whether to stay or leave. The pattern across those conversations is the one I want to set out here. The traditional law firm partnership is not failing because AI is better than lawyers. It is failing because the cost structure of legal service delivery is changing under it, and the legal-form vehicle most firms still operate cannot absorb the change.

The article works through the four reasons the partnership form cannot hold, what the strategic clock looks like in real years, the cautionary post-mortem of the firm that converted but failed to make the transition stick, and what each of three audiences should now do.

Share LevelUP Strategies

Why this matters now: the AI capex shift

For most of legal history, the partnership form worked because the only thing legal needed to fund was lawyers. The expensive thing was the associate’s time. You billed the time, the client paid, and at year end the partners distributed what was left. The partnership distributed everything because there was nothing else to invest in. Property leases were modest. Practice management software was a rounding error. Knowledge management was a partner with a good memory.

That is no longer the cost structure of competitive legal service delivery. The expensive thing now is the platform: the data, the workflow systems, the integrations, the engineering team that builds and maintains them, and the multi-year contract with the foundation-model provider that powers it. AI moves legal from a variable-cost business (associate hours) to a fixed-cost business (platform engineering). The partnership form was never tested against fixed cost structure at this scale, because legal never had fixed cost structure at this scale.

I was reminded of this earlier this week listening to the Legal Innovation Spotlight, in which Ted Theodoropoulos and Scott Kveton (co-founder of CaseMark), walked through what the post-AI legal landscape actually looks like from the vendor side. Scott put the structural problem directly. “It doesn’t behoove [LLPs] to make investments in technologies to advance their firm.” The phrasing is mild. The argument is not. He is saying that the partnership form structurally punishes investment, and that the firms still operating under it are competing in a capital market with one hand tied behind their back.

The capital concentration on the other side of that fight is the part most legal commentary still understates. Harvey has raised at a valuation around $11 billion. Legora is at roughly $5.5 billion. Freshfields recently announced a direct partnership with Anthropic, bypassing the legal-tech intermediary layer entirely. The investors funding those vehicles are not asking for a modest return next year. They are asking for an out-sized return after a decade of compounding the platform. A partnership distributing every penny of profit to its partners every twelve months won’t be able to keep up. It’s not even in the same race.

That is the macro picture. The structural details are what make it deterministic.

The 4 reasons the law firm partnerships cannot hold

1. The partnership punishes saving

A partnership is a pass-through entity for tax purposes. The tax authorities treat profits as belonging to the partners the moment those profits are earned, whether or not the partners actually take the money home. So if the firm decides to keep £200,000 inside the business to fund a platform investment, every partner pays income tax on their share of that £200,000 in the year it is earned, even though it never reached their bank account.

Compare this to a corporate-form business. The company pays corporation tax on what it retains and pays the post-tax remainder out as dividends if and when it chooses. The double layer of tax is real. Critics of corporate form make much of it. What that critique misses is that the double layer is the price for being able to hold a balance sheet. A C-corp can own assets bigger than this year’s salary bill. A partnership cannot, because every pound the partnership tries to save incurs a cost to the partners personally that year.

The household analogy makes this concrete. Consider two families with identical incomes. The first splits its entire income at year end and spends it. The second family retains a portion inside a household account, pays tax on it, and uses the savings to renovate the kitchen and put down a deposit on a second house. Twenty years later, one family owns three properties. The first family still earns the same income and still spends every penny.

The partnership has been the first family, by structural necessity, for as long as the form has existed. It did not matter when there was nothing to invest in. It matters now.

2. The partnership cannot pay engineers in equity

The people who build AI-native legal platforms typically are not lawyers. They are engineers, designers, data scientists, and platform operators, and they do not work only for a share of cash bonuses. Often, they work for shares in the businesses they help build. The reason is straightforward. The expected upside on shares in a successful technology business is the only compensation that competes with what the same engineer could earn at Anthropic, Stripe, or a venture-backed startup with a four-year vest.

A traditional law firm cannot give shares to a non-lawyer. The regulator does not permit it. So the firm offers a salary, the engineer compares the salary to a competing offer from a non-law firm with shares on top, and the engineer takes the competing offer. This is not a culture problem. It is a market-clearing problem.

The Alternative Business Structure regimes solve it directly. The United Kingdom enabled non-lawyer ownership of legal businesses through the Legal Services Act 2007, with the Solicitors Regulation Authority licensing the first ABSs in 2012. By 2022/23, 1,202 ABSs had been licensed, representing 13 per cent of all law firms in the UK. Arizona Supreme Court Rule 31.1 introduced the first ABS regime in the United States in 2020. Utah’s regulatory sandbox, also from 2020, performs similar work via different machinery.

The most consequential entry under any of these regimes happened in February 2025, when the Arizona Supreme Court approved KPMG Law US as the first Big Four legal services entity in the country. KPMG global revenue in 2025 reached $39.8 billion, with tax-and-legal services growing at roughly 8 to 10 per cent year on year. A multidisciplinary parent of that scale can absorb a decade of capex losses building out its US legal arm. No partnership has equity partners with comparable patience.

The talent gap compounds annually. Every year a firm cannot offer equity is a year its ABS-eligible competitors hire ahead. Five years of that is a generation of platform engineers locked in elsewhere.

3. The partnership cannot raise outside money

A partnership cannot sell shares to outside investors. It cannot list on a stock exchange while operating as a partnership. The available capital is what the partners themselves contribute, plus what the firm can borrow from a bank, and bank borrowing is constrained because partners are personally liable for partnership debts. The result is a capital ceiling set by partner risk appetite and partner cash flow.

Set that ceiling against the AI vendors. Combined Harvey and Legora valuations sit somewhere north of $16 billion. The capital they have raised has already been deployed, is being deployed, and will continue to be deployed against the same legal-spend pool the law firms compete for. If the maths of those vendor valuations is to be made whole, they need to capture a non-trivial share of the global $1.1 trillion legal spend. Twenty per cent. Thirty. Possibly more. The exact share is debatable. The fact that they need to take share, aggressively, is not.

Viewed through that lens, the Freshfields-Anthropic deal is the more consequential story than the Claude-for-Legal webinar. A Magic Circle firm went directly to a foundation-model provider, bypassing the legal-tech vendor layer entirely. The deal-by-deal dynamic on that direction of travel is straightforward. The frontier labs need to demonstrate that they can serve sophisticated legal buyers at scale. Sophisticated legal buyers want to access the underlying capability without paying a wrapper margin. Both sides have an incentive to disintermediate the middle layer. Whichever side wins on a given deal, the partnership form on the buy-side is not on equal negotiating terms with the well-capitalised labs.

Either firms restructure to access patient capital and build their own platforms, or vendors and frontier labs absorb the work directly. Those are the two endpoints. There is no third.

4. The partner-track economics no longer make sense

The compact between an associate and a law firm has always been simple. Grind for seven years. Make partner. Take a slice of an enterprise that grows over the next twenty. The compact assumes a growing pie. It collapses if the pie shrinks.

A bright associate four years in can now look at three options. The first is to keep grinding for a partnership share in a firm that may be measurably smaller in ten years. The second is to leave for a venture-backed legal-tech startup with shares that vest in four. The third is to start a new firm, structured from day one as an ABS or corporate entity, with engineers as co-founders and investor capital behind it. The second and third options are starting to win on simple maths, particularly for associates already exposed to AI work and able to assess the trajectory clearly.

Scott Kveton’s framing for this in the afore-mentioned podcast was the “midlife-crisis associate” at three to five years post-qualification. He is right that this is the founding pool for the next generation of legal businesses. This generation already has the technical literacy. They have the legal context. They no longer have a deal worth staying for. The associate exodus that results is not a culture problem. It is a value-of-the-deal problem, and culture cannot fix maths.

Once associate departure rates pass a certain threshold, partnership succession becomes uncertain. The lateral market reprices accordingly, the bench thins, and the firms most exposed to AI-substitutable work hit the inflection first.

That moment has already arrived in some firms. The denial period that typically follows is the first six to twelve months of leadership convincing itself the departures are anomalous.

Read my related article on the 3 types of businesses inside a law firm, and the implication on where AI will have the greatest impact.

The strategic clock

The unfortunate property of structural conversion is that it takes most of a decade. Knights plc is the cleanest data point we have. The firm decided in 2012 to convert from a seven-partner partnership into a corporate entity with five shareholders. They listed on the London Alternative Investment Market in 2019 at a £103.5 million valuation, raising £50 million. It took seven years from the structural decision to prepare for the public-capital event.

CEO David Beech’s framing of the timeline, in interviews around the IPO, was that “the best way to get ready to list is to ditch the partnership model well in advance.” The phrasing is precise. The conversion is not the listing. The listing is the conversion clearing.

On the available evidence, the gap between deciding to restructure and being structurally ready to compete is somewhere between five and seven years.

Knights is one data point and worth treating with appropriate caution. The direction of the inference is robust nonetheless. There is no observed case of a partnership-to-corporate conversion happening cleanly in two years.

Working backwards, then. A firm that wants to be AI-native by 2032, with a balance sheet that can fund the platform investment that AI-native delivery requires, needs to begin structural conversion no later than 2027. Firms that decide in 2030 will not have a balance sheet capable of capitalising the work until 2035-2037. Firm owners not taking decisive action on this timeline are bought, merged, or quietly wound down.

2027 is now the deadline that matters. 12-18 months’ time. Closer than the typical strategic-planning horizon of most managing partners.

The cultural transition trap

The Knights timeline is the optimistic version. The DWF Group timeline is the cautionary one.

DWF went public in March 2019, listing on the London Stock Exchange main market and raising £95 million at a valuation just under £370 million. They were the first major UK law firm on the main market. The IPO price was 122 pence. The share price peaked around 143 pence in early 2020 and then fell, eventually trading around 65.5 pence, a loss of approximately 54 per cent for IPO investors. Inflexion took DWF private in 2023.

The post-mortem on what went wrong is the part of the story that matters. From the published commentary at the time of the take-private, DWF’s equity partners faced significant reductions in annual remuneration in order to release sufficient profit for shareholders. The structural conversion happened. The compensatory and behavioural rewiring did not. Partners who had spent careers extracting maximum cash every year were asked to forgo distribution in exchange for compounding equity. That is a shift few managed to sustain. The legal entity changed shape. The economic behaviour inside it did not.

The lesson here is not that structural conversion fails. It is that conversion is two transitions, not one, and the second is harder than the first. The legal restructure is a project. The cultural restructure is a generational change in how partners think about what they own. Firms that wait until 2030 to start the legal restructure will then face the cultural restructure on top of it, with AI platform investment running in parallel, in a market in which their best associates are already leaving and their largest clients are already procuring around them. There is no time to do those three things at once.

The DWF outcome is the consequence of telescoping the two transitions. Begin the structural restructure early enough that the cultural transition can run sequentially rather than in parallel, and the chance of stabilising into a genuinely capitalised vehicle is materially better. Begin late, and you have a corporate entity full of partners who still behave like partners. The market then prices the firm accordingly, which is exactly what it did to DWF.

What this means for you

What follows are three audience cuts. Each is short by design. Each should function as a self-contained brief that can be read alone or sent to a peer to raise awareness and spur debate.

For managing partners

The question to model on a whiteboard, honestly, is “what would our balance sheet look like as a corporate entity by 2030, if we began restructuring this year?” The output of that exercise dictates which of four roles your firm plays in the second half of the decade. You are an acquirer, a merger candidate, an acquisition target, or a deliberately small ceiling-by-design specialist boutique. Doing nothing is also a choice – and you’d be choosing to become an acquisition target. The window to be an acquirer is closing, because the firms that began converting in 2018 to 2020 are entering their listing window now. Their balance sheets are being capitalised while yours is still being fully paid out and distributed.

For mid-career associates

The partner track was a deal between two parties. Both parties have to honour the deal for it to make sense. Your side is the seven-year grind. The firm’s side is a growing pie. If you cannot see a growing pie at the firm in front of you, the rational move is to look at restructured firms that can offer equity, or to join the next-generation legal businesses being founded right now. Many of those will be founded by people in exactly your position, with engineer co-founders and outside investor capital. The structural transition is also a founding moment. It happens once a generation. It is happening now.

For corporate general counsel

The structural form of your panel firms is now a procurement question, even if you have not been asked to think of it that way. The right question to ask each of your panel firms is straightforward. What proportion of last year’s profit was retained inside the firm? What is the firm’s three-year capex plan for AI-driven legal service delivery? Who are the engineers and platform operators on the team, and what is the equity vehicle that holds them? Firms that cannot answer the third question will be slower, more expensive, and structurally less capable. You will notice within five years, probably three. The cost of switching panel composition now is materially lower than the cost of switching in 2030.

Close

The traditional law firm partnership is not failing because AI is better than lawyers. It is failing because legal service delivery is moving from a labour business to a capital business, and a partnership that distributes profits annually cannot fund a capital business. The Alternative Business Structure regimes are the route across. The window to begin the restructure is now narrower than the time it takes to complete the restructure. That is the whole story.

The structural question is the one that matters. Technology, talent, and culture follow from the answer to the structural question, not before.

If you are a UK law firm exploring a partner exit – partial or full – the ABS route is often the cleanest path. The regulatory structure allows outside capital, allows non-lawyer equity, and allows you to retain the firm’s name and client relationships whilst restructuring the partnership economics.

If that is relevant to where you are, we should talk. I advise law firms on exactly this structural question. No sales pitch, no obligation. Just a conversation about your available options.

→ Book a call with me

The action you take about your firm’s model dictates every strategic move that follows.

Please consider sharing this with a colleague you think would benefit.