MANIFESTO LAWAGENT · EDITION 6 · July 2026

Invisible Efficiency: why your firm doesn't know a case's real margin

By LawAgent Team · Institutional editorial analysis · 7 min read · 25:22 min podcast
Invisible Efficiency: why your firm doesn't know a case's real margin

Executive Summary

Ask a managing partner how much the firm billed last year and he'll answer with precision. Ask the real margin of one specific case — what was left after deducting everything that case cost — and the answer, at most firms, will be silence. Artificial intelligence brings a property the human hour never had: its cost is exactly traceable. A precise cost was the missing piece for the return on technology investment to stop being an article of faith and become a calculation.

A restaurant that doesn't know which dishes on the menu turn a profit and which run at a loss wouldn't survive a quarter. Most firms have operated this way for decades — they close the year in the black, conclude that everything is fine, and never see the composition behind the aggregate. Revenue is billed with precision; margin is measured in the dark. This edition is about the difference between the two, and about the instrument that finally dissolves it.

The firm flying blind

There's a number that sums up the economics of the legal profession: roughly 90% of fees still flow through hourly billing, according to the Thomson Reuters Institute's 2026 State of the Legal Market Report. The figure is American — in Brazil, hourly rates, retainers, fixed fees, and contingency fees all coexist — but the lesson holds for every model, because the problem isn't how you charge, it's what you measure. The billable hour measures the input (time spent), not the outcome (what was left over). And the gap between the two is larger than it seems: realization — the fraction of nominal value a firm actually collects — fell from 84.5% at the end of 2021 to 81.9% by mid-2024 at U.S. firms. Of every dollar "produced" by the rate card, close to a fifth evaporates between the work and the cash register. And almost no firm can say which cases caused that loss.

Worse: across a portfolio of dozens of cases, some turn a profit and others run at a loss, and the aggregate hides the composition. A prestigious corporate client can consume far more work than the contract pays for, quietly carried by the fees of smaller, more profitable clients. The firm believes it is serving its most important client; in reality, it is subsidizing them.

Running a firm without knowing margin per case is, quite literally, flying blind — and flying blind works until the day it doesn't.

What the cloud taught us: FinOps

In the 2010s, tech companies went through an accounting scare that the legal profession is only now starting to live through. Migrating to the cloud, they traded a fixed, predictable cost for a variable, diffuse one — and the bill arrived inflated and faceless, with no indication of which team or client had generated the expense. The answer was a new discipline, FinOps (a contraction of Finance and DevOps): stop treating technology as a vague, opaque overhead line, and attribute every dollar spent to the business unit that generated it. According to the FinOps Foundation, which standardizes the practice, it's the model that "maximizes the business value of technology" and "creates financial accountability" through data.

The key concept the cloud popularized is called unit economics. Instead of asking "how much did we spend in total?", you ask "how much does each unit of what we produce cost, and how much does it return?" The legal equivalent is immediate: cost and margin per case, per client, per type of filing. And it's worth clearing up a misunderstanding: FinOps was never a pair of scissors for cutting spending — it's a pair of glasses. Sometimes it recommends spending more, deliberately, where the return is high.

▶ WATCH THE FULL DISCUSSION

An introduction to legal FinOps and measuring the return on artificial intelligence.

Traceable cost and the ROI that finally shows up

ROI is return on investment: net gain divided by the cost of what was invested to obtain it. Applying it to legal technology has always run into a double obstacle — neither the gain nor the cost was measurable at a firm.

The gain hides a paradox. If a tool cuts the time to draft a filing in half, under the hourly model that gain doesn't turn into profit: fewer hours mean fewer hours to bill, and therefore less revenue. Efficiency, in the world of the hour, is self-cannibalizing — the better you get, the less you charge. The return showed up on the spreadsheet as a loss, and the real gain stayed invisible, hidden in hours that stopped existing. Hence this edition's title.

Cost, on the other side, was a fog: self-reported human hours, rounded, logged from memory; expenses allocated by crude averages. This is where AI changes the nature of the problem. Unlike the human hour, the machine's work leaves an exact meter — every search, every draft consumes a precise amount of processing, logged the instant it happens and attributable to a specific case. The variable cost of production stops being fog and becomes a number.

An architecture built to be auditable is, as a bonus, an architecture built to be costed. Security and management draw from the same well: the trail.

The traceability that Edition 3 described as the foundation of reliability — every claim attributable to a source — is the same traceability that, seen through a financial lens, attributes every cent to a case. With a precise cost in the denominator, the ROI of legal technology finally gains the two things it needed to become a calculation: a gain that can be isolated and a cost that can be measured.

Real margin by case and client

Once variable cost becomes traceable and revenue is already known, calculating the margin of each case becomes routine — the job of a financial control tower. An example makes the math concrete (illustrative numbers, not actual data). Two clients pay the same fixed monthly fee of 8,000 reais. Client A's cases cost 3,000 reais to produce — a 62% margin. Client B's, thornier, cost 7,800 reais — a 2.5% margin. Under the old model, they'd show up on the same spreadsheet line, treated as equals. With per-case visibility, two uncomfortable truths become evident: Client B is nearly running at a loss, and Client A's surplus had been silently subsidizing it.

What this visibility unlocks is decision-making, not automatic judgment. The firm can renegotiate scope, rethink how certain cases are produced, or knowingly keep a loss-making client for strategic reasons — as long as the decision is a conscious one. Seeing the margin is not the same as chasing only high-margin cases: margin informs, it doesn't dictate. A loss-making case can be a strategic relationship, an institutional cause, pro bono work done out of duty and by choice. The value of visibility lies in making conscious a choice that today is made in the dark.

🎧 LISTEN TO THE FULL PODCAST

The full discussion on margin, traceable cost, and ROI in the legal profession.

From visibility to pricing

The ability to measure margin has a consequence that touches the most sensitive point in legal economics: how you charge. Alternatives to the hour — flat package fees, contingency, subscription — are the subject of a future edition in this series and won't be developed here. What matters is the sequence: measurement comes before price. Moving from the hour to a flat fee means taking on a risk — that the work costs more than the agreed price. Taking on that risk without knowing your own cost is gambling, not pricing. Without a floor of real cost, there's no intelligent ceiling on price.

There's an elegant symmetry with Edition 5: that one argued for measuring the professional by contribution, not by the hour; this one argues for pricing the work by value, not by the hour. It's the same move on two planes — career and price — and both depend on ceasing to measure the input and starting to measure the outcome.

Limits and responsibility

The legal profession is not a spreadsheet. Reducing cases to margins would commodify a profession with duties that don't fit into a fraction — a fair concern. The answer: margin informs, it doesn't dictate, the way a thermometer reports temperature without forcing any decision. A firm that starts accepting only what's profitable will have understood the tool backwards. The metric serves judgment; it doesn't replace it: the number is an input to the partner's decision, never an automatic verdict — the same human oversight that governs the entire architecture applies here (Edition 3).

Communication and fees. Margin visibility is a management capability, not a promise that the firm will make more money, and nothing here suggests using cost to increase fees — fee-setting has its own boundaries under the OAB Code of Ethics, which prohibits undervaluing professional fees and improper client solicitation. What's on record is what Provimento No. 205/2021 requires: informative, sober, free of commercialization. Confidentiality and the LGPD (Brazil's data protection law): cost data derives from the firm's own cases; logical isolation between clients (tenant isolation) is non-negotiable, and a firm's financial data belongs to it and stays locked to it.

The firm that doesn't know its real margin per case isn't lacking profit — it's lacking instruments. And flying blind works until the day it doesn't.

LawAgent · Editorial thesis

The efficiency of legal technology has, for decades, been invisible: real, but drowned in a model that only knew how to count time. That's not a law of nature — it's a lack of instruments, and the instrument now exists. When part of production comes to be done by a machine whose work leaves an exact meter, cost becomes a number, margin becomes a calculation, and return becomes evidence. The difference between managing with the data and without it is the difference between a pilot flying by instruments and one flying by looking out the window, hoping the weather stays clear.

This content is for informational and educational purposes only and does not constitute legal advice.

If this analysis was useful to you or your firm, share it with other senior partners. Manifesto LawAgent's content is designed to circulate among Brazil's top legal professionals.

EDITORIAL NOTE

LawAgent Team

Institutional editorial analysis

Content produced by the LawAgent team with the support of artificial intelligence tools, based on its own research. LawAgent is the legal AI companion designed for senior partners, specialized boutiques, and in-house legal departments.

Upcoming editions

The next edition of the LawAgent Manifesto will be published soon.

Take this edition in PDF.

Executive version ready to circulate internally at your firm.

Download PDF

(PDF in Portuguese)

Want to see how LawAgent applies these ideas in the day-to-day of Brazilian law firms?

Subscribe to LawAgent-ai→