The Economics of AI: Rethinking Law Firm Pricing and Profitability

Welcome to Attorney Intelligence, the weekly newsletter from PointOne where we break down the forces reshaping legal from the inside out.

This week, I want to move past the question everyone is asking: Is AI killing the billable hour?

The question we should be focusing on, the one that decides firm profitability, sits on the other side of the P&L. What happens to a law firm’s finances when a large share of its work becomes extremely cheap and fast to produce?

The answer is not automatically “the firm becomes more profitable.”

Around 75% of partners do not know their fair value, and 54% do not understand the relationship between pricing and profitability. A firm cannot set a defensible fixed or value-linked fee if it cannot identify which matters, staffing models, clients, and practice groups make money.

What law firm costs reveal

Operational work supports revenue

Administrative and operational teams make legal service delivery possible, but their time is generally treated as overhead rather than billed directly to clients.

When AI improves these workflows, the firm can expand capacity or reduce costs without directly reducing revenue. If revenue holds, margins improve.

Characteristics of well-positioned firms

Lawyer compensation is a cost, but billable lawyer time is also what the firm sells.

When AI reduces the hours required for a matter, it lowers the labor input needed to deliver the work. Under hourly billing, however, it also loses the revenue attached to those hours. The firm becomes more efficient, but most of the matter’s revenue disappears with the hours.

AI investment therefore has two different financial impacts:

• Automating operational work can increase capacity and reduce overhead while preserving revenue.

• Automating billable work can reduce both cost and revenue unless the pricing model changes.

Law firms spend most of their money on people. Labor consists of between 50% and 65% of a law firm’s total costs. Overhead—the cost of running the firm beyond directly producing legal work—accounts for around 30% of total costs at the largest firms to 39% at smaller firms, leaving the residual as operating profit. In 2025, law firms increased lawyer headcount by 2.9%, salaries by 8.2%, technology spending by 9.7%, and knowledge-management spending by 10.5%, suggesting that the average firm is investing in AI without reducing its overall cost base.

This creates the worst possible combination: firms are adding labor and technology costs while retaining a pricing model that reduces revenue when lawyers become faster.

The simplest approach is to divide the work into two broad categories:

Commoditized work. Commoditized work does not need to be easy or unimportant to become a commodity. It only needs to become predictable enough that another provider can perform it to an agreed standard for a published price.

Non-commoditized work. Non-commoditized work cannot be purchased at a fixed fee because it is directly tied to rare experience, judgment, or accountability.

The goal is not to hide the efficiency from the client. It is to divide the benefit fairly: the client pays less, but the firm wins the client business and earns more than it would under hourly billing.

How to identify the commoditized work

Around 61% of in-house legal leaders say they continue sending some work to law firms mainly out of habit, not because they have decided that the law firm is the best provider. The clock is ticking on the reliance on habit. AI has fundamentally challenged taking the comfortable path, pushing clients to innovate in-house to stay ahead of their own competitors.

Until recently, law firms generally had to optimize for two of three competitive advantages: speed, price, and accuracy. AI has not eliminated those tradeoffs, but it has brought delivering all three a closer reality. This is particularly true for predictable tasks such as large-scale document review, first-pass contract analysis, and routine entity formation.

Identifying commoditized work is fairly simple. Could a competent third party perform this work to an agreed standard at a set price without relying on the personal judgment of your partners?

If the answer is yes, the work has already become or is becoming commoditized.

Commoditized work does not equal easy, unimportant, or free from risk. A diligence review can still determine whether a transaction proceeds. A regulatory filing can still have serious consequences. The work may remain essential, but once its delivery becomes predictable and the market is saturated with capable providers, it loses the power to command a premium.

What remains valuable, and why?

Certain types of legal value still cannot be purchased for a simple published price.

Precedent and experience

An AI tool may identify 400 issues in a diligence review, but an experienced partner knows which 12 could affect the purchase price or determine whether the deal should proceed.

That judgment comes from recognizing patterns across similar matters—how companies, courts, opposing counsel, regulators, and transactions have handled comparable issues—and connecting those outcomes to the current strategy. Today, this experience is scarce because it often lives inside individual partners’ heads.

Firms can capture and organize this knowledge to share it across teams, but the subjective judgment built through experience remains difficult to codify or replace.

Reputation

A firm’s reputation can influence a matter before the legal battle even begins.

The opposing side may negotiate differently when facing a firm known for taking cases to trial, defending particular claims successfully, or maintaining credibility with a specific regulator. That reputation affects what the other side believes the client is prepared to do—and therefore which strategies, demands, or concessions it considers credible.

In these situations, the client is buying not only the firm’s work, but also the market response its involvement creates.

Risk allocation

Negotiation is not simply about winning every point.

Experienced lawyers understand:

• Which issues truly matter

• Which points can be conceded

• The best order in which to make concessions

• How a particular counterparty is likely to respond

• Which contract structures are likely to create disputes later

That judgment can change the financial and legal outcome of a matter.

Accountability

The law firm remains professionally responsible for its work, even if AI produced the first draft. The software does not stand behind the advice. The firm does.

Clients are not only buying documents or research. They are also buying the confidence that a qualified professional reviewed the work and accepts responsibility for it.

However, firms must be able to prove the value they claim to provide.

Many clients cannot reliably determine whether one law firm’s technical work is better than another’s. Clients therefore judge firms using things they can observe, such as responsiveness, communication, and reliability.

A firm may genuinely be faster, more accurate, or more strategic. But if the client cannot see evidence of that difference, the market will price the firm as though the difference does not exist.

Price the work according to what the client is buying

The shift away from hourly billing does not require law firms to price every matter around outcomes. It starts with recognizing that a single matter may contain different types of work—and that each should be priced differently.

Commoditized work should be priced against the client’s next-best alternative, whether that is an ALSP, an in-house team, a smaller law firm, or a legal technology solution. Work that defines the firm’s value can support a staged base fee with a smaller payment tied to a clear milestone, such as signing, regulatory approval, financing, or closing.

Investment banks offer a useful model by separating payment for the work performed from payment tied to a successful transaction. Law firms cannot control every factor that determines an outcome, so most of their fee should remain guaranteed. But a smaller milestone payment can reflect the value at stake without placing the firm’s entire fee at risk.

AI will make routine legal work faster and easier to compare across providers. As that happens, firms will need to understand which parts of their work are becoming commoditized, where their judgment creates distinct value, and what it actually costs to deliver both. The firms with the clearest view of their own work will be best positioned to price it.

Legal Bytes

Harvey builds its own brain. The $11bn legal-AI company introduced Tenet, its first proprietary model for legal work, giving it a way to lower costs and reduce its dependence on outside model providers like OpenAI and Anthropic (Business Insider).

Digital twins come for lawyers. Twin1 emerged from stealth with a $20m seed round to build AI versions of professionals that learn from their emails, meetings, documents, and workplace systems. Linklaters, Orrick, and Dechert are already using the platform (Business Wire).

Elevate buys the workflow layer. Elevate acquired Lupl, the legal project-management platform built to coordinate lawyers and AI tools across complex matters, adding native connections to Claude, Harvey, and Copilot (Elevate).

Value-based pricing still depends on time data

At PointOne, our read is that the next decade rewards firms that can measure and replicate what drives productivity, margin, and profitability, starting with the data underneath every decision: time.

Book a demo to see how PointOne helps firms understand where their time and resources are allocated.

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

Katon