AI Has Fundamentally Changed My Litigation Practice

I've been practicing commercial litigation for almost ten years - ten years of my professional life measured in one-tenth of an hour fractions. Six-minute increments.

For almost two years now, I've been steadily integrating AI into my practice. Today, I use enterprise-grade AI tools daily. They have augmented my capacity and changed how I plan and execute the workflow involved in managing files.

THE EFFICIENCY IMPERATIVE

Non-lawyers (and many lawyers) are incredulous that we still bill in 0.1-hour increments. There are constant calls for reform. But for professions dealing in complex, non-routine tasks, no better model has emerged. Clients are paying for access to our cognitive bandwidth and the expertise that comes with it.

Tracking time so closely can be grueling. It creates constant pressure to be efficient and accountable for every 0.1. The reality of a growing practice - multiple complex files, each with their own deadlines, deliverables, and personalities - makes these increments feel necessary, but no less relentless.

Not every task carries the same value, but every minute is billed the same. That has always made me uncomfortable. The true measure of value lies not in each unit, but in the cumulative outcome for the client. Which is why AI has been such a game-changer: it helps me shift more of my limited time toward the tasks that add the most value.

AUGMENTED CAPABILITIES

I often tell new lawyers that developing capacity means filling "buckets of competency" - research, drafting, advocacy, communication, strategy, and judgment. Practicing law is like playing a multi-dimensional board game with half a dozen rule books: statutes, contracts, common law, procedure, evidence, and professional obligations. Each decision means weighing competing considerations.

What ties these buckets together is one skill: knowing what is relevant. That was the first lesson of law school, and it remains the most important. Anyone can collect facts or cite cases. The lawyer's value lies in spotting what matters, framing it clearly, and guiding the client forward.

Technology has always helped us process information more efficiently. But generative language models are different. They don't just store or retrieve; they augment the lawyer's own ability to synthesize, summarize, and test ideas.

I think back to one of the first lawyers I trained under, who dictated all his materials. With documents spread out in front of him, he'd pause, think, press record, and dictate entire drafts out loud. He could do in an hour what took me six. I was in awe of the efficiency.

A decade later, I can do something similar. I dictate the outline of my thoughts, and AI synthesizes, refines, and clarifies them instantly. It gives me that same speed and precision, while enhancing the depth of my analysis.

PRACTICE MANAGEMENT

I draft thousands of pages of material every year. One of the hardest skills to learn is estimating how long it will take to prepare written work. For years, I thought of drafting in two stages. Getting to an 80% draft - assembling facts, evidence, and documents into a coherent first pass - took half the time. Closing the gap from 80 to 100% - editing, refining, and perfecting - took the other half.

That framework has changed. AI compresses the first stage dramatically. It can pull facts from evidence and correspondence, organize them coherently, and provide references in a fraction of the time. The second stage has sped up too. I can query the file, check for inconsistencies, and generate alternative framings in minutes.

AI hasn't replaced my judgment. It multiplies my capacity to exercise it. The result: more time on the part of my work that matters most - analyzing, advising, advocating.

It has also changed the rhythm of drafting. Like most lawyers, I've agonized over phrasing, spending hours trying to distill complex points into a single sentence. Now, once I know the components, I can generate variations instantly. The focus has shifted from how do I say this? to what do I need to say?

Day to day, the way I use AI looks mundane, but it matters:

These are the building blocks of practice: identifying, synthesizing, packaging information. They also strain a lawyer's capacity. Having a tool to accelerate them, while still leaving me responsible for verification, has fundamentally changed how I work.

UNDERSTANDING THE RISKS

Adopting new technology isn't without risk. Overreliance on AI has already been well publicized. Long before I used it in practice, I was reading about algorithmic systems. In The Master Algorithm, the author draws a simple analogy: you don't need to know how to build an engine to drive a car, but you do need to understand the steering wheel, accelerator, and brakes. AI is the same.

At their core, these tools are advanced prediction machines. Trained on vast amounts of text, they generate language by predicting the next word. They use neural networks - layers of algorithms that recognize patterns - and a mechanism called "attention," which lets the model weigh which parts of a prompt are most relevant. That's why the way you frame a question matters so much. The attention mechanism is the steering wheel.

Like any prediction system, AI has blind spots. It can overfit, clinging too tightly to patterns it has seen before, or underfit, missing nuance. That's why outputs can look convincing but still be wrong. Garbage in, garbage out. The quality of the prompt determines the quality of the result.

Sometimes prompting feels like shaking a Magic 8 Ball. As a kid, I thought I could trick mine into giving me the answer I wanted. Sometimes it worked, often it didn't. AI can feel the same. You can chase the perfect prompt and hit a dead end. But over time, you build intuition. The more you use it, even for small, low-stakes tasks, the better you get at driving it safely from point A to point B.

On security and confidentiality, I see the same concerns many do. But the rules aren't fundamentally different from cloud computing. The platforms we use don't train their models on client data, and the information is secured to the same standard already applied to sensitive legal work.

The bigger risk is us. Lawyers are busy. Deadlines loom. The temptation to take outputs at face value without checking them is real. But the safeguard is the same as it always was: verify the source information for everything.

LOOKING FORWARD

That is how I've come to see AI in my practice. It doesn't replace the buckets of competency I've spent ten years developing. It augments them. It multiplies my capacity but doesn't change the underlying obligation to exercise judgment, know the file, and stay rooted in the evidence.

This technology has made me more of the lawyer I wanted to be: spending more time on judgment, strategy, and advocacy, and less on the grind of moving a mountain of documents six minutes at a time.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.