By Jeffrey T. Donner, Esq.

September 9, 2026

There is a persistent misunderstanding about what an experienced lawyer is supposed to know. The misunderstanding usually takes this form: if a lawyer has practiced for twenty-five or thirty years, particularly in a specialized field, why should that lawyer need to spend several hours researching the law? Shouldn’t an experienced lawyer already know the answer?

The premise sounds reasonable only until one considers what lawyers actually do. No competent lawyer, however experienced, knows every case, every statutory amendment, every contractual provision, every insurance endorsement, every procedural variation, and every factual permutation that may arise in a substantial piece of litigation. One of the differences between an experienced lawyer and an inexperienced one is often that the experienced lawyer has learned how dangerous it is to assume that yesterday’s answer necessarily controls today’s case.

Law changes. Contracts differ. Insurance forms are revised. Courts distinguish earlier decisions. Statutes are amended. A seemingly minor factual variation can change which precedent controls. Sometimes an endorsement buried deep in a policy changes an analysis that initially appeared straightforward. Sometimes an appellate decision issued only a few years earlier has refined what lawyers had previously treated as settled law. Research therefore is not evidence that the lawyer does not know the law. Properly conducted, it is evidence that the lawyer understands the difference between knowing a field and knowing the answer to the particular case sitting on the desk.

Experience Does Not Eliminate Research

An experienced construction lawyer may understand the basic architecture of Florida construction law. An experienced insurance lawyer may be thoroughly familiar with occurrence-based coverage, business-risk exclusions, resulting-property-damage principles, subcontractor exceptions, allocation issues, and the distinction between liability and insurance coverage. A trial lawyer may have decades of experience with summary judgment, expert testimony, burdens of proof, damages, preservation issues, depositions, and the strategic consequences of discovery. None of that means the lawyer can responsibly advise a client without reading the actual contract, the actual insurance policy, the actual endorsement, the actual engineering report, the pleadings, and the cases that govern the particular issue.

The same point is illustrated by specialization itself. Florida recognizes board certification in construction law because construction practice is sufficiently complicated to warrant a formal specialty credential. A board-certified construction lawyer has demonstrated substantial involvement in the field, completed specialized continuing legal education, undergone peer review, and passed a written examination. Certification does not mean that the lawyer somehow carries the entire current body of Florida construction law in memory. It means the lawyer has demonstrated the knowledge and experience necessary to recognize the issues, understand their significance, and perform the work required to answer them correctly.

That distinction applies across sophisticated fields of practice. A tax lawyer still checks the statute and regulations. An appellate lawyer still researches whether an issue was preserved under the particular record. A coverage lawyer still reads the policy. A construction lawyer still reads the new construction cases. An experienced commercial litigator still determines which authorities actually control the dispute in front of the lawyer.

Indeed, experience often creates additional research questions rather than eliminating them. A sophisticated lawyer does not merely ask, “What is the rule?” The lawyer asks whether a later decision narrowed the rule, whether a particular endorsement changes the analysis, whether a factual distinction that looks trivial is actually dispositive, whether the burden of proof changes at a particular procedural stage, and whether an attractive quotation from an appellate opinion was part of the holding or merely dicta. Experience tells the lawyer which questions may matter. Research supplies the answers.

What Large Law Firms Traditionally Sold Was Organized Intellectual Labor

I learned the traditional model firsthand. Earlier in my career, I practiced at firms including Holland & Knight, Gunster, and Podhurst Orseck. In a large-firm litigation environment, difficult legal questions were routinely addressed through a division of labor that was both intellectually rigorous and expensive.

A partner or senior lawyer would identify an issue and assign it to an associate. The associate might spend substantial time reading cases, tracing a line of authority, examining statutes, comparing conflicting decisions, reviewing documents, and writing an internal memorandum. The senior lawyer would then review the research, challenge the analysis, request additional work where necessary, consider how the authorities fit the facts, and determine how the issue should affect the client’s litigation strategy.

There was nothing unusual in that environment about an associate spending several working days researching a difficult legal question and preparing an interoffice memorandum. I remember periods as a young lawyer when I billed ten-hour days researching and writing those memoranda, and that time was billed to clients. The partners supervising the work were not assigning research because they lacked knowledge of their fields. They were using a rational division of professional labor. The associate performed the labor-intensive first-stage analysis; the senior lawyer supplied experience, quality control, strategy, and judgment.

That model remains recognizable in sophisticated commercial and construction practices today. Large firms maintain teams of lawyers at different levels of seniority because substantial litigation involves many different tasks, and it generally makes little economic sense for the most senior lawyer to perform every first-pass assignment personally. A junior or midlevel lawyer investigates the issue and organizes the authorities; a more senior lawyer determines what the research means for the case.

Historically, however, the client paid for both layers. A difficult issue might generate substantial associate time followed by additional senior-lawyer review. Even when the associate’s hourly rate was materially lower, the total cost of reaching the final answer could be substantial.

I would not suggest that every difficult question historically required thirty, forty, or fifty billable hours, much less that any particular law firm would bill that amount today. Billing depends on the assignment, the record, the lawyers, and the client. But I do know from firsthand experience that substantial blocks of associate time were routinely devoted to precisely the sort of research, document review, organization, and memorandum writing that technology can now accelerate.

Artificial Intelligence Compresses the First Layer of Work

Artificial intelligence is beginning to alter that traditional division of labor. Used properly, AI can perform portions of the work that associates historically performed manually. It can identify potentially relevant cases, locate provisions buried in lengthy documents, organize factual chronologies, compare authorities, generate issue lists, summarize competing arguments, and produce a first-pass explanation of a complicated legal framework.

That can save real time. On some assignments, it may eliminate several hours that twenty-five years ago would have been spent by an associate locating authorities, organizing them, and preparing an initial memorandum. On a sufficiently complicated assignment, the savings may be considerably greater. In that sense, AI can remove or compress part of the first layer of billable labor in a traditional law-firm staffing model.

But that is very different from saying that AI eliminates most of the lawyer’s work. In my experience, it frequently does not eliminate even half of it.

The senior lawyer still has to read the source documents. The lawyer still has to examine the policy, contract, pleading, expert report, deposition testimony, statute, regulation, or other governing material. The lawyer still has to verify the authorities independently, read the important opinions, determine whether they remain good law, understand their procedural posture, distinguish holdings from dicta, identify adverse authority, and decide whether a purported factual analogy is genuine or superficial.

The lawyer then has to do something AI cannot accept professional responsibility for doing: synthesize the authorities with the actual facts, exercise judgment, determine what additional evidence is necessary, anticipate the opponent’s arguments, assess litigation risk, and counsel the client about strategy, expense, settlement, discovery, and trial.

Artificial intelligence therefore does not replace the senior lawyer’s research obligation. It can remove some of the preliminary labor necessary to get the lawyer to the important part of the research.

That distinction is particularly important because an AI-generated legal memorandum, however polished, is not something responsible counsel can simply forward to a client. If an AI system identifies a case, the lawyer must confirm that the case exists and actually says what the system claims it says. If it identifies a policy provision, the lawyer must read that provision in the policy itself. If it proposes an argument, counsel must determine whether the argument survives scrutiny under the governing authorities and actual record.

In my own practice, that frequently means using AI as one research tool and then returning to traditional legal research platforms and primary authorities to verify and deepen the analysis. The lawyer remains responsible for the answer. Technology does not change that.

The Senior Lawyer Still Has to Read Everything

The most important point about AI-assisted legal work is also the least glamorous: the lawyer still has to sit down and read.

Suppose an AI system can accomplish in minutes a preliminary task that once required a junior lawyer five or ten hours. That is a genuine efficiency gain. But if the resulting material takes an experienced lawyer six hours to read, verify, compare with the authorities, understand, synthesize with the facts, and convert into reliable advice, those six hours remain real professional work.

That is not evidence that the technology failed. It means the technology eliminated the lower-level portion of the assignment rather than the judgment-intensive portion.

Indeed, the more experienced the lawyer, the less likely the lawyer should be to accept an AI-generated answer at face value. An experienced litigator knows that apparently simple legal propositions often contain qualifications. The jurisdiction matters. The procedural posture matters. The burden of proof matters. The precise policy or contractual language matters. An appellate court may have announced a broad proposition while resolving the case on a much narrower ground.

This is one of the reasons I believe artificial intelligence can improve quality rather than merely lower cost. If technology reduces the amount of professional time spent mechanically locating and organizing material, the experienced lawyer can devote a greater percentage of the available time to testing the answer, considering adverse arguments, evaluating the evidence, and thinking about what should actually be done.

The final work product can therefore be better even when AI does not dramatically reduce the senior lawyer’s own billable time.

Artificial Intelligence Also Reduces Clerical Labor

There is another category of work that AI reduces, and it has little to do with legal research: the mechanical production of words.

The evolution of legal technology shifted a considerable amount of clerical work onto lawyers themselves. Earlier generations of attorneys often dictated correspondence, memoranda, and briefs to secretaries or transcription personnel. By the time my generation entered practice, much of that model was disappearing. Lawyers increasingly typed and revised their own work and performed tasks that once would have involved substantial support staff.

I graduated from law school in 1999, when that transition was already well underway. Like many members of Generation X, I learned to type in school on an actual typewriter. By the time I became a lawyer, typing my own work was not unusual; it was simply part of practicing law.

For more than twenty-five years, lawyers of my generation have therefore spent substantial time physically producing words in addition to deciding what those words should say. I type very quickly, but even fast typing consumes time and attention. Every letter, memorandum, motion, outline, and revision still has to be physically produced.

AI is beginning to reduce some of that mechanical burden. A lawyer can dictate a complicated thought in conversational language, reorganize a draft, compare alternative formulations, summarize background material, or create a first-pass document without manually typing every sentence from scratch. The lawyer still has to review and take responsibility for every word that ultimately goes out under the lawyer’s name, but less professional time has to be spent functioning as a typist.

There is some historical irony in that development. Earlier generations benefited from substantial secretarial support. Lawyers of my generation largely absorbed much of that work themselves. AI may now return some of that time to substantive analysis.

What Is a Reasonable Legal Fee?

The economics of this discussion also illustrate why evaluating legal fees solely by looking at the number of hours on an invoice can be misleading.

Florida’s Rules of Professional Conduct recognize that a reasonable legal fee involves considerably more than arithmetic. Rule 4-1.5 directs lawyers to consider, among other things, the time and labor required; the novelty, complexity, and difficulty of the questions involved; the skill necessary to perform the service properly; the rate customarily charged for comparable legal services; the significance and amount involved in the representation; the results obtained; and the lawyer’s experience, reputation, diligence, ability, and efficiency of effort.

That framework makes sense because legal services are not commodities measured solely by time.

Suppose one lawyer charges $800 per hour and correctly resolves a difficult issue in four hours, while another lawyer charges $400 per hour and requires eight hours to reach the same correct answer. The mathematical cost is identical. The second lawyer is not necessarily providing inferior value simply because the assignment took longer. What matters is whether the lawyer possessed the competence to identify the issues, find and understand the controlling law, protect the client procedurally, develop the necessary evidence, make the appropriate arguments, and exercise sound professional judgment.

The same principle works in the opposite direction. A lawyer charging a very low hourly rate is no bargain if the lawyer misses a dispositive case, overlooks an exclusion in an insurance policy, allows a deadline to expire, fails to obtain necessary discovery, or pursues a theory that cannot succeed. Legal services have economic value only if the work is competent.

Law is not mathematics in the sense that competent lawyers cannot promise that a particular argument will produce a particular judicial result. Judges disagree. Juries disagree. Witnesses surprise everyone. Facts develop differently than expected. No responsible trial lawyer can convert litigation into an equation that guarantees an outcome.

But there is an important sense in which legal practice does contain an objective component. A competent lawyer should identify the controlling authorities. A competent lawyer should understand what those authorities hold. A competent lawyer should recognize important procedural deadlines. A competent lawyer should know which facts have to be proved and should make the arguments reasonably available under the governing law.

There is therefore a threshold question that matters more than the hourly rate: can this lawyer competently handle this matter?

If the answer is no, the lawyer should not be hired merely because the hourly rate is attractive. If the answer is yes, then the economic question becomes whether the lawyer can provide the necessary level of service at a reasonable total cost relative to the amount at stake and the value being pursued.

That is a much more useful inquiry than comparing hourly rates in isolation.

Why Six or Eight Hours of Research May Be Highly Efficient

Clients sometimes look at a time entry showing six or eight hours of research and conclude that the lawyer must have been inefficient. In sophisticated litigation, that conclusion may be exactly backward.

Consider what a serious research and analysis assignment may entail. The lawyer may have to identify the controlling framework, locate and read the leading appellate decisions, determine whether they remain current, distinguish favorable and unfavorable authority, compare those decisions with a particular contract or insurance policy, determine whether an endorsement changes the analysis, study the factual record, identify missing evidence, and convert the resulting legal conclusions into practical advice.

If an experienced lawyer can perform all of that carefully in six or eight billable hours, that may be a highly efficient result.

The proper comparison is not between eight hours and zero. The relevant question is how much total professional labor would otherwise have been necessary to produce an equally reliable answer.

Based on my own experience in large firms, an issue that an experienced lawyer can now investigate with AI assistance and analyze in several hours might previously have generated a substantial associate research assignment, followed by additional senior-lawyer review. The associate might have spent one, two, or several working days locating, reading, organizing, and summarizing the authorities before the senior lawyer began reviewing the work. The client could then receive an invoice containing both layers of professional time.

AI can compress the first layer. It does not eliminate the second.

The client today does not see those avoided hours. The client sees only the invoice that exists, not the invoice technology prevented. This can create a peculiar paradox for efficient small-firm practice: the better technology works, the easier it becomes to underestimate the professional labor that would previously have been required.

Suppose a traditional firm assigns a partner and one or more associates to investigate a complicated issue. The lawyers collectively devote thirty hours to research, analysis, internal consultation, and preparation of the final product. Now suppose an experienced small-firm lawyer uses artificial intelligence to accelerate the first-stage research and organization, personally verifies the important authorities and governing documents, and reaches the same reliable strategic answer in eight or ten billable hours.

If the quality of the analysis is equivalent, the relevant economic fact is not merely that the small-firm lawyer billed eight or ten hours. It is that the client obtained the necessary professional result without paying for the additional layers of staffing.

This is where hourly rate and efficiency interact. A sophisticated small firm may already charge a materially lower hourly rate than a major commercial firm while simultaneously eliminating substantial associate time through technology. If the lawyer nevertheless delivers a comparable professional result, the client benefits in two ways: fewer total hours and a lower rate for the hours actually billed.

That proposition necessarily depends on the premise that the smaller firm gets the legal work right. If it does not, the lower bill is largely irrelevant. But when an experienced small-firm lawyer can deliver the required quality, the economics can be compelling.

Legal Value Is Not Necessarily the Same Thing as the Legal Bill

There is an additional point that hourly billing can obscure. Legal work creates value by advancing or protecting the client’s substantive interests, not by generating units of time.

Consider a hypothetical construction dispute involving an $800,000 loss. Suppose competent legal work ultimately produces a substantial recovery that otherwise would not have occurred. The value created for the client may bear relatively little relationship to whether one lawyer needed seventy hours and another needed ninety hours to accomplish the result, particularly if the lawyers charged materially different rates.

This does not mean lawyers should be indifferent to efficiency or that the amount in controversy makes any fee reasonable. A lawyer should not perform $100,000 worth of unnecessary work simply because $800,000 is at stake. Time, labor, customary rates, complexity, experience, ability, efficiency, and results all matter.

But the opposite mistake is equally serious. A client should not focus so narrowly on avoiding several thousand dollars of necessary legal work that the client materially impairs the ability to pursue a claim worth hundreds of thousands of dollars.

The rational inquiry is expected value. If a client has a plausible claim with a substantial potential recovery, legal expense should be considered in relation to the amount at stake, the probability of success, the likely cost of reaching a resolution, and the alternatives available. A client may rationally decide that a claim is not worth pursuing. But that decision should be based on economics and risk, not on the assumption that sophisticated litigation can be prosecuted without meaningful legal expense.

Hourly billing sometimes obscures that perspective because clients naturally focus on the bill that arrives today rather than the value the representation may create years later. A $7,500 invoice is tangible and immediate. A potential $400,000 recovery several years in the future is uncertain. That psychological difference is real, but it does not alter the expected-value analysis.

The Economics Become More Difficult When a Small Plaintiff Faces an Institutional Defendant

The difference between efficiency and unrealistic expectations becomes particularly important when an individual, small business, condominium association, or other noninstitutional plaintiff has a substantial claim against a major corporation or insurer.

The institutional defendant does not approach litigation as an occasional expense. Litigation is part of its operating environment. Large insurers and corporations maintain claims personnel, internal lawyers, document systems, experts, litigation guidelines, outside counsel, and established procedures for defending lawsuits. Their outside lawyers may work at negotiated rates substantially below ordinary commercial rates, but the institution can still authorize the hours necessary to defend the case.

If counsel needs to review the policy, counsel reviews it. If the lawyers need to research an exclusion, they research it. If depositions are necessary, they take them. If expert testimony is required, experts are retained. If a substantial dispositive motion requires substantial attorney time, the institutional defendant has a mechanism for evaluating and paying for that work.

A smaller plaintiff may have a claim worth several hundred thousand dollars and nevertheless experience every legal invoice as money being removed from reserves, operations, savings, or another immediate need. That concern is understandable. A corporation and an individual or condominium association do not necessarily experience a $10,000 litigation expense in the same way.

But that difference does not alter what litigation requires. A plaintiff cannot realistically demand the benefits of sophisticated litigation while refusing to fund the professional work necessary to obtain them.

A client is always free to decide that a claim is not worth pursuing. Clients should establish budgets, question substantial expenses, demand explanations, and reassess cases as evidence develops. Those are responsible decisions. But there is a difference between controlling litigation expense and imposing a budget so restrictive that the lawyer cannot do the work necessary to prosecute the case effectively.

That distinction matters particularly when the opponent has institutional resources. A claimant cannot reasonably expect to prevail against a sophisticated insurer merely by sending correspondence and hoping the company voluntarily pays. If the dispute becomes genuine, the claimant’s lawyer must understand the policy, anticipate the defenses, develop the evidence, use appropriate experts where needed, conduct discovery, and prepare the case in a manner demonstrating that the claim can actually be litigated.

Institutional defendants understand the difference between a demand supported by a developed case and one that is not.

The Answer Is Staged Litigation, Not Unlimited Litigation

None of this means that a client should give a lawyer an unlimited budget. The sensible response to litigation expense is disciplined, staged litigation.

Lawyer and client can identify the next factual or legal question that materially affects the value of the case and authorize sufficient work to answer it. One phase may focus on causation and damages. Another may seek records identifying responsible participants or available insurance. Another may involve a focused expert analysis. Once those pieces are developed, counsel may prepare a supported demand and determine whether the defendant is willing to resolve the dispute before substantially more expensive litigation occurs.

At each stage, counsel should be able to explain what the next work is intended to accomplish, its anticipated cost, and how the resulting information will affect the decision whether to proceed. If the evidence materially weakens the case, the client can stop. If the evidence strengthens it, additional expenditure may become economically rational.

That is responsible litigation management. It acknowledges uncertainty without pretending uncertainty can be eliminated before money is spent.

Staged litigation nevertheless has a floor. A task that reasonably requires four hours cannot honestly be compressed into one merely because the client would prefer a smaller invoice. The cases still must be read. The contract still must be analyzed. The policy still must be examined. The evidence still must be understood. The expert’s conclusions still have to be reconciled with the governing legal standard.

Efficiency can reduce the cost of those tasks. It cannot responsibly eliminate them.

Research and analysis are also not ancillary to litigation. Frequently they are the litigation. Hearings, depositions, mediations, and trials are the visible events, but the machinery underneath consists of document review, factual investigation, legal research, expert development, witness preparation, discovery planning, motion practice, and strategic judgment.

A lawyer who arrives at a hearing without having done that work may produce a smaller invoice. The client may eventually discover that the apparent savings were expensive.

AI Can Also Increase Billable Work

There is another side to generative technology that deserves attention: AI does not always reduce lawyers’ time. Sometimes it creates additional work.

Clients now have access to many of the same generative tools lawyers use. That can be beneficial. A client may use AI to organize facts, prepare a chronology, identify questions, or summarize documents before consulting counsel.

But generative systems also make it possible to produce enormous quantities of superficially sophisticated material almost instantly. A client can generate fifty, one hundred, or even 150 pages of purported legal analysis containing cases, statutes, constitutional arguments, technical terminology, and confident conclusions. Some of it may be useful. Some may be irrelevant. Some may be subtly incorrect. Some may rely on authorities that do not support the propositions for which they are cited.

The lawyer still has to read it.

If a client asks counsel to evaluate 150 pages of AI-generated legal analysis, the fact that the material took minutes to generate does not mean it takes minutes to review. Counsel must determine what matters, verify the authorities, identify errors, explain why apparently attractive theories may fail, and redirect attention toward the issues that actually affect the case.

In that circumstance, AI has increased rather than reduced professional time.

This illustrates a fundamental distinction that extends well beyond law: generative AI has dramatically reduced the cost of producing words. It has not correspondingly reduced the cost of determining whether those words are correct.

In legal practice, that difference matters enormously.

AI Is Changing What Young Lawyers Are Worth

The traditional role of the junior associate is changing as well. Young lawyers are not disappearing, and sophisticated firms will continue to need them. But some of the economic justification for traditional junior-associate work is changing rapidly.

For decades, one important function of a junior litigation associate was supplying large quantities of intelligent labor. A young lawyer might receive a complicated research problem and spend ten or twenty hours locating cases, reading them, organizing them, and preparing a memorandum. The process also trained the associate. Through repetition, young lawyers learned how courts reasoned, how precedents fit together, and how apparently minor factual distinctions could change an outcome.

AI can now perform parts of that mechanical process much faster. The junior associate whose primary value is the ability to spend an entire day locating and summarizing cases therefore operates in a different professional environment from the associate of twenty-five years ago.

That does not mean young lawyers are less valuable. It means their valuable skills are changing. The associate of the future must know how to formulate the right research question, use technological tools intelligently, verify primary authorities, identify hallucinations, locate missing adverse authority, distinguish holdings from dicta, and transform machine-assisted research into analysis on which a partner, client, or court can responsibly rely.

AI may therefore raise the standard for associates rather than eliminate them.

A technologically sophisticated young lawyer may be capable of accomplishing in several hours what once took considerably longer. That associate can become more valuable, not less, by combining technological fluency with the traditional intellectual discipline of legal research.

The same observation applies at the other end of the experience spectrum. Some younger lawyers may adopt new technology more readily because they have fewer established habits to change. Yet the greatest potential value of AI may arise when technological capability is combined with decades of substantive experience.

An experienced lawyer already knows what reliable analysis looks like. The lawyer can recognize when an AI system has identified the wrong issue, relied on an irrelevant procedural posture, confused related doctrines, or offered an attractive argument that is unlikely to survive contact with the record. The experienced lawyer understands that an impressive quotation may be dicta, that materially different policy wording can destroy an analogy, and that a factual distinction the system treated as minor may determine the outcome.

Technology supplies speed and organizational capacity. Experience supplies skepticism, context, and judgment. Used together, they can produce better legal work.

Clients Should Demand Efficiency Without Confusing Necessary Work With Waste

Clients are entitled to scrutinize legal bills. Lawyers should be able to explain what they did, why they did it, and how the work advanced the client’s interests. There is nothing inherently valuable about an hour merely because a lawyer recorded it. Unnecessary research remains unnecessary, and duplicative staffing should be controlled.

Technology should reduce legal expense where technology can legitimately reduce it.

But clients should not confuse necessary intellectual work with waste merely because the work occurs at a desk rather than in a courtroom. Reading controlling appellate decisions is not waste. Studying the governing insurance policy is not waste. Comparing an expert’s conclusions with the governing legal standard is not waste. Anticipating an opponent’s defenses before filing suit is not waste. Determining what evidence must be developed before making a demand is not waste.

Those activities are the substance of litigation.

An experienced lawyer may perform them more efficiently than an inexperienced lawyer. AI may reduce the necessary time further. Neither experience nor technology eliminates the obligation to perform them correctly.

This kind of efficiency has the potential to expand access to sophisticated legal representation. A client with a substantial commercial dispute may be unwilling or unable to finance a traditional large-firm staffing model involving several lawyers at different billing rates. The same client may be able to retain an experienced small-firm lawyer who uses technology to eliminate some of those layers while personally performing the work requiring senior judgment.

Historically, large institutional firms possessed an enormous information-processing advantage. They had large numbers of associates, libraries, research personnel, litigation-support departments, databases, and lawyers specializing in adjacent subjects. A solo lawyer could possess considerable trial experience and still lack the institutional capacity to process information on the same scale.

Modern technology is narrowing that gap.

A sophisticated small-firm lawyer using modern legal-research platforms, AI-assisted analysis, electronic discovery and filing systems, cloud-based litigation tools, and appropriate outside experts can now perform some work that once required a much larger institutional platform. There remain cases whose scale genuinely requires large teams; one lawyer does not become twenty lawyers simply by acquiring better technology. But for a significant category of commercial disputes, the intellectual-capacity advantage of the largest firms is materially smaller than it once was.

That development should benefit clients. It creates the possibility of obtaining senior-level representation without necessarily paying for an entire law-firm pyramid.

The necessary tradeoff is that clients must understand what senior-level legal work actually entails. The lawyer can eliminate waste. The lawyer can use technology. The lawyer can avoid unnecessary staffing. The lawyer can divide the case into rational phases. What the lawyer cannot ethically eliminate is the analysis necessary to determine what the law is, what the facts establish, and whether the client’s position can be proved.

The Real Promise of AI in Litigation

Artificial intelligence does not make legal work free. Nor does it ordinarily eliminate most of the lawyer’s billable time. What it can do is remove a layer of preliminary labor, reduce clerical work, improve organization, accelerate issue spotting, and allow experienced counsel to spend a greater percentage of professional time on the work that actually requires professional judgment.

Serious litigation involves adversaries, and adversaries create work. Evidence must be gathered. Documents must be reviewed. Witnesses must be prepared. Experts may be required. Depositions may have to be taken. Motions must be researched and written. Judges must be persuaded. Opposing arguments must be answered. No technology eliminates those realities.

What technology can change is how much human labor must be devoted to getting the lawyer to the point where professional judgment becomes decisive.

A lawyer who once needed an associate to spend two days identifying and organizing the relevant cases may now be able to identify the important authorities much more quickly and spend the lawyer’s own time reading, testing, and applying them. A lengthy insurance policy can be searched and compared in ways that would previously have required many hours of manual review. A large factual record can be organized rapidly enough that the lawyer spends more time thinking about what the evidence means and less time simply locating it.

That should reduce the client’s cost. It may also improve the final work product.

But greater efficiency should not create a new misconception that competent analysis should therefore require little or no billable time. If an experienced lawyer spends six hours researching and analyzing a complicated issue with the assistance of modern technology, the six hours should not automatically be viewed as six hours too many. They may represent fifteen, twenty, or more hours that the client never had to pay for under the traditional staffing model.

Even when AI produces the first draft, the experienced lawyer still has to sit down and read it. The lawyer still has to verify it. The lawyer still has to understand it. And ultimately, the lawyer—not the software—has to decide whether the answer is right and what the client should do about it.

That may be the simplest way to understand the technology’s actual role in modern law practice. Artificial intelligence can help identify issues, find authorities, organize information, and produce a preliminary analysis. It cannot relieve the lawyer of responsibility for determining whether the analysis is correct.

The future of legal practice is therefore unlikely to be a choice between lawyers and machines. The more important distinction may be between lawyers who understand how to use technology to enhance their judgment and lawyers who do not.

The best use of AI is not to eliminate research, associates, or lawyers. It is to reduce the mechanical cost of getting from a difficult legal problem to the point where an experienced lawyer can do what experienced lawyers have always been paid to do: determine what matters, decide what can be proved, assess what the opponent will argue, protect the client’s procedural rights, and advise the client what to do next.