Welcome to Attorney Intelligence, where we break down the biggest advancements in AI for legal professionals every week.

This week, I’m thinking about the AI law firm model. I spent last week working alongside Claude in Cowork mode: Leveraging it to research, pressure-test arguments, and analyze large datasets. What used to take me a full day took a morning. And here’s the thing: The output was better. Not because the AI is smarter than me. Because the AI let me operate at a level I couldn’t reach on my own.

That’s the shift most people are getting wrong about legal AI. It’s not about replacing lawyers. It’s about what happens when deep expertise meets a tool that can operate at the speed and scale that expertise deserves. An experienced litigator who knows exactly what to look for in a medical record doesn’t need AI to think for her – she needs it to cover ground she physically can’t.

AI is a multiplier for experts. And legal is one of the best applications on the planet for it, because legal work is knowledge-intensive, language-heavy, and deeply dependent on the kind of judgment that only comes from experience.

So the real question isn’t “which AI tool should we buy?” and it isn’t “how many associates will we need in five years?” The real question is bigger and more exciting: How do we rethink the law firm itself to capture this opportunity? The firms that define the next era won’t be the ones that reacted best. They’ll be the ones that built first. And that means asking the hard questions:

How do we hire and train differently? How do we plan financially around a model where technology amplifies every expert in the building? How do we finally productize legal services – not as a consumer play like Rocket Lawyer, but as the way sophisticated firms serve sophisticated clients?

For over a century, the law firm has run on the same formula. Partners own the equity and bring the clients. Associates provide leverage. Revenue scales with billable hours. The era changed and it’s the most exciting time we’ve seen. The firms that see this as an opportunity to rebuild from the ground up, rather than a problem to solve only with a software purchase, will define the next generation of legal services.

I believe the conversation starts with the data that drives every firm’s revenue engine: time. Here’s how I am thinking about it.

1 - Why now is the moment for rethinking the law firm operating model

Legal tech is now moving at light speed

Here's what makes this moment different from every previous wave of legal technology: The speed of building has fundamentally changed.

The founder of Casetext spent years building legal AI with traditional Natural Language Processing (NLP), years of work that never quite reached what lawyers needed it to do. When large language models arrived, he accomplished in months what had taken years, shortly before Thomson Reuters acquired the company for $650 million. Harvey went from its founding to a $8 billion valuation in three years. Legal tech startups raised $4.3 billion in 2025, a 54% increase from the year before. The time to reinvent legal services is now.

In-house expertise is not enough to ship a product

As the cost of building software dives toward zero, the operative question is not “who has the best technology?” but rather "who understands the problem deeply enough to build the right thing?"

Big Law has the in-house expertise needed to innovate. Orrick has an internal lab where 200 lawyers spend roughly 6,000 hours a year on innovation. Freshfields opened a lab in Berlin with data scientists co-creating tools alongside clients. But while these units have produced real tools (see. Clerky, conceived by two Orrick attorneys), they faced the structural constraints inherent to the partnership model. To ship their product, successful innovators first had to exit the firm.

The partnership model isn’t designed to commercialize technology; it’s designed to commercialize time. Those are fundamentally different business models.

But productizing legal work is possible

Rocket Lawyer raised $288 million, serves over 25 million users, and is valued north of $1 billion. It proved that standardized legal work could be productized and sold directly to consumers. But even at that scale, it didn't touch law firm economics: Document creation was automated, but the infrastructure that determines how a firm prices, staffs, or delivers the work behind that document were unchanged.

UpCounsel tried a different angle, a freelance marketplace connecting clients with independent lawyers, promising flexibility and lower costs. The matching got better, proving that distribution could be reimagined, but the costs associated with the work remained consistent.

Atrium is maybe the most instructive case. They raised $75.5 million to bundle a law firm with proprietary software and a subscription model starting at $500/month. The subscription idea was right, clients loved predictable pricing. But two things killed it: The AI didn't exist yet to meaningfully compress legal labor costs, and sinking capital into building proprietary software from scratch was the wrong bet when the real bottleneck was the intelligence layer itself. The thesis wasn't wrong. The timing was. That AI exists now.

Diversifying supply chains in legal services

Document automation proved legal work could be productized. Marketplaces proved distribution could be reimagined. Atrium proved clients wanted alternative pricing. Innovation labs proved the talent and expertise lives inside firms. No single attempt could change the operating model on its own, but together, they demonstrated the building blocks. The firms that have the courage to now build the foundation for a new operating model will reinvent how legal work is accomplished.

Crosby, backed by Sequoia and Index Ventures, charges $400 flat per contract review with a 58-minute median turnaround time. Justpoint launched as the first AI-native mass tort firm under Arizona's alternative business structure rules, embedding AI directly into case intake and pricing. These companies didn't come out of law firms. They couldn't have, given the structural barriers. They are ultimately able to ship a product because their internal operations are targeted toward delivering that product, not billing for time.

Here's the insight most people miss: Law firms don't have to compete with Crosby. They could be Crosby's customers. A firm that can't productize NDA review internally could outsource it to a platform that does it in under an hour at a fraction of the cost. That means Crosby represents the possibility of specialized supply chains that previously did not exist within the legal industry.

2 – A blueprint for the law firm of the future

Buying better software is no panacea

The data is brutal: 88% of business transformations fail to achieve their original ambitions when it comes to implementing tech innovation. Failed digital transformations cost $2.3 trillion per year globally, wasting time and resources on consultants, software products — and still facing low adoption.

That's because most of what firms are buying right now is “vitamins”, not “medicine”. Vitamins make you feel slightly better about what's already there. Medicine fixes something broken.

AI law firms will have a wholly new architecture

Hybrid ownership will give diverse stakeholders a piece of the pie

If the core asset of a firm becomes its proprietary workflows, AI systems, and data infrastructure, not just rainmaking partners, then ownership logic has to follow. Arizona removed its ban on non-lawyer ownership entirely. Utah's regulatory sandbox is live. KPMG got approved to practice law in Arizona.

Alternative business structures allow hybrid ownership: lawyers, technologists, capital providers under one roof. The firms willing to share ownership with the people building their most valuable assets will attract better talent and move faster than the firms that don't.

Better data will enable a rigorous approach to culture change

New ownership only works if it's paired with new infrastructure. Organizations that invested in culture change alongside technology saw 5.3x higher success rates than technology-only approaches. So how do law firms approach culture change rigorously?

I believe that reengineering the core infrastructure starts with good data. Understanding how time is captured, how work is priced, how profitability is measured, this is the engine that either enables or blocks every other transformation a firm attempts. You can't make informed decisions about what to productize, what to automate, and where your experts add the most value if you don't have visibility into how the work actually flows today.

Visibility into culture allows leaders to reconsider the revenue model. You can't hire differently, build differently, and own differently while still billing by the hour. When the primary input shifts from human time to technology, pricing has to shift with it. The firms that understand which services to productize and which engagements will remain relationship-driven will have a structural advantage in the evolving market.

New definitions of legal talent

Think about finance. Not everyone in finance is a CPA. You have controllers, FP&A analysts, revenue ops managers, treasury specialists, financial engineers. Dozens of distinct roles, each with a finance component but a different skill set. In contrast to other mature industries, legal still acts like it's lawyers and "everyone else."

The AI law firm model needs legal engineers, professionals at the intersection of product, law, and solution architecture. Tech paralegals who can configure AI workflows. Legal operations specialists who think in systems, not cases. The American Bar Association is already writing about how legal engineers are reshaping modern law firms. Salaries range from $95,000 to $149,000. The role exists at the intersection we describe in our piece on The Rise of the Legal Engineer.

Firms that harness new supply lines will win

The AI-native law firm isn't a faster version of what exists. It's a different economic organism, one where the pyramid is flatter, the infrastructure is smarter, the talent is more diverse, and the ownership reflects who actually creates the value.

You can't run AI-native workflows on 1990s infrastructure. The processes, the incentives, the org chart, the revenue model, they all have to move together or none of them move at all. This isn't a technology problem. It's a systems problem. And the firms that solve it will be positioned to both serve clients directly and plug into the emerging supply chain of productized legal services.

Every previous attempt to change the model, document automation, marketplaces, hybrid firms, innovation labs, moved the industry forward. None of them had the technology to finish the job. Now we do. The firms that see this moment for what it is, not a threat but the most exciting opportunity in a generation, will build what comes next. And it starts with the data that powers every decision: time.

Legal bytes

Here are the latest updates in legal tech and AI:

AI startup hopes to revamp government IT systems — AI startup Cognition has launched Cognition for Government, an AI coding and software engineering platform aimed at migrating agencies’ outdated computer systems to modern programming languages (Bloomberg Law News).

Anthropic scales back safety committments due to competitive pressures — AI company Anthropic is softening its core safety policy to stay competitive with rivals like OpenAI, xAI, and Google (Wall Street Journal).

Perplexity’s legal battle with Dow Jones heats up — Perplexity, an AI-powered web search engine, filed another motion in its copyright lawsuit again Dow Jones, arguing that the newspapers “cherry-picked” AI-generated responses to validate their case (Business Insider).

Curious about transforming your firm using time data?

Book a demo to see how PointOne can help turn time data into clear insights, increasing visibility into firm culture and operations.

Thanks for reading and I'll see you next week,

Katon