This is the law firm of the future

Judith del Cuadro-Zimmerman and I recently presented this to members of the DC Bar. Watch the full session above, or read the highlights below.

Over a decade ago, Judith started the first version of her law firm with two things: a course on how to start and run a law firm, and a whole lot of enthusiasm.

Then the clients came. And with them, the invoices.

Judith was billing clients with a Word document. She'd type up line items by hand, one invoice at a time. With a handful of clients, no big deal. But the firm grew, and before she knew it, invoicing was eating days of her month. She'd come home to Chris (her boyfriend at the time, now her husband) and vent that she hadn't billed a single hour. All she'd done all day was send invoices.

Chris said, "I think we can solve this together."

One weekend later, they'd built the firm's first billing system in Excel. Days of invoicing became half a day. Judith used that spreadsheet for years. Actual years.

It got her sanity back. It helped her land more clients, bill more hours, and keep growing. Over time, the firm graduated to more sophisticated tools like Clio.

But there was still one massive problem: everything still depended on Judith. She was the only person who understood the systems she'd cobbled together. She was the only person who really knew what was happening on each case.

That's about when we met.

From pedaling harder to switching vehicles

I'm not an attorney. I'm an entrepreneur and an advisor (and a proud Canadian dialing in from Austin, Texas). My job is to help businesses grow and become sustainable — able to stand on their own.

When I started advising Judith's firm, DCZ Legal, it was crystal-clear to me that she couldn't keep pedaling the proverbial bike any harder. She needed to switch vehicles altogether.

So we got to work. Today, DCZ Legal has surpassed $1 million a year in billings. Judith has influenced legislation. She closed a deal with one client that brought multiple cases into the firm at once. And for the first time, she can see a path to owning a law firm that doesn't depend entirely on her.

Last year, I laid out for Judith what I saw as the law firm of the future: a firm that embraces the best of AI while protecting itself from the downside. She was all in. So we started a new management company together: Law Firm Partners.

To be clear, these are two different entities. DCZ Legal is the law firm. Law Firm Partners is a management company we co-own — it's not a law firm. More and more, Law Firm Partners runs DCZ Legal's back office: marketing, intake and sales, finance, operations, and tech. That frees Judith and her attorneys to focus on being attorneys.

The results so far? 30% to 90% time savings on the tasks that support the firm.

And here's the part I'm most excited about. The AI tools behind those savings aren't ChatGPT. They aren't Claude, Perplexity, or Harvey, either. They're tools we built from the ground up — compliant with the American Bar Association, secured at an enterprise-grade level, and producing remarkable accuracy. Oftentimes better accuracy than doing the work by hand. (No knock on anyone. Any human doing the same task over and over for hours will make the odd mistake.)

It's not really about AI

Before we go under the hood, I want to name something Judith said that stuck with me.

Tech isn't the point. Tech is what powers the firm to do what she wanted to do in the first place: help people protect what they've built and provide for the people they love — especially people in vulnerable places.

Technology is just the support system for being a great attorney.

The problem every firm is living with

Here's what onboarding a new client used to look like at DCZ Legal. My guess is it looks a lot like yours.

A new client signs the legal services agreement. That triggers a cascade for support staff:

• Open a new matter in Clio.
• Set up the client in Google Drive.
• Send out questionnaires.
• Let the supervising attorney know a new client is coming on board.

Multiple staff members. Multiple platforms. Information everywhere. Even with a sophisticated tech stack, nothing was integrated.

And it gets worse.

Say a matter goes quiet for three months. Then opposing counsel sends a demand letter. Now the attorney has to figure out where they left off — digging through Clio, Google Chat, and old emails, then asking staff what they did last and where the documents are. Basically, the firm had to hold a team meeting every time it picked up a case that was a couple of months old.

All of that work is confusing. All of it takes time. And none of it is billable.

When Judith and I zoomed out — looking at DCZ Legal and at the other firms I advise and Judith's many attorney friends — we kept seeing the same three issues: simplicity, compliance, and work quality.

So the question we set out to answer was this:

How can a law firm get the massive AI advantage that's already here — and getting bigger by the day — while simplifying its tech, staying compliant and safe, and keeping work quality high?

The answer: a firm headquarters

A firm headquarters is your law firm's own custom-built software that holds all the information that's currently scattered across a dozen different platforms.

Think about a fictional client named Margaret Alvarez. She calls and says, "I've got a new email address." In most firms, someone now has to update her in the practice management software. Then the project management software. Then make sure Gmail uses the new address. Then update the phone system.

That's happening in law firms across DC, across America, and across the world. Every day.

With a firm headquarters, there's one Margaret Alvarez record, and it feeds everything else.

The core of any firm's productivity comes down to five areas:

• Messaging — attorney to paralegal, paralegal to clerk, and so on.
• Contacts — clients, prospective clients, and related parties.
• Matters — the cases themselves.
• Project management — a production board showing where every matter stands. (Most firms don't have this, and would benefit greatly from it.)
• Task management — a front page that says, "Here's what you need to do today." (Same story.)

In our build — Law Firm Partners HQ — all five live in one place. No separate Slack or Google Chat. No separate Asana or monday.com. No separate task manager bolted onto a separate practice management tool.

That alone is a game changer. But the real unlock comes next.

Private AI: your walled garden

Once all your information lives in one place, you can use what we call a private AI. (If you're more technical, you'd call it a private LLM.)

Here's why that matters. When you use ChatGPT, Claude, Perplexity, Harvey, or any similar subscription tool, your data has to leave your firm and go to that third party. Different vendors make different claims about security. Some train on your data. Some let you turn that off. Some don't.

A private AI works differently.

If you're over about 35, you'll remember installing Microsoft Word from a CD in the '90s. (North of 40? Maybe a floppy disk.) Word lived on your computer, and that was that.

Most people don't realize you can do the same thing with AI. You can install an AI model directly on your own computer — or, more advanced, on your own cloud server.

Now you've created a walled garden. Your client notes, your documents, your spreadsheets, and your AI all sit inside the same four walls. No data goes out. No data comes in.

And here's the catch: a private AI loses its potency if it has to leave the walled garden to do anything, and your data loses its security if it has to leave, too. That's why the headquarters and the private AI go hand in hand.

This is absolutely the future of law firms. In our opinion, it's only a matter of time before it becomes the standard.

Four customizations that weren't possible before

Here's where it gets even better. When you own your headquarters, you can build customizations around your firm's exact needs — things no off-the-shelf tool can do.

Your practice area may look nothing like Judith's elder law and fiduciary practice. That's okay. Don't get caught up in the details. Notice the principle. I guarantee your firm has its own versions of these, and then some.

1. Mail processing: 15 hours a week down to 1.5

Because DCZ Legal serves as a fiduciary for people in the community, it receives mail for roughly 50 other individuals. Sorting it had become a Herculean task.

The old process: one staff member opened and scanned everything into Google Drive. Someone renamed every file to match the firm's naming system. Then all seven team members dug through the same mail folder, trying to figure out who needed to take next steps, and sent tasks to each other in Clio. Then someone billed for every piece of mail.

Seven people. Two or three platforms. 15 hours a week — just to sort mail.

Today, the headquarters automates the renaming, routes tasks to the right team members, and captures the billables. It takes 1.5 hours a week, overseen by one staff member instead of seven.

That's 90% less time. Without losing a single billable.

I often hear the worry that AI will take human jobs. I see it the other way around: humans have been doing robot jobs for a long time. Let the robots do the robot jobs, and free the humans to do higher-level work. The relief on team members' faces every time we bring them a new automation is remarkable — and there's no shortage of meaningful work waiting for them. (More on that in Will AI Take My Job?)

2. Call logging that finds the next actions for you

DCZ Legal has an in-house receptionist. Every call meant figuring out who was calling and which matter it related to, taking notes, logging those notes in the practice management software, sending tasks to teammates, messaging people on Google Chat, and scheduling callbacks. One person juggling three platforms to capture one call — with plenty of room for mistakes along the way.

Now the phone system (Dialpad) sends every call straight into the headquarters. For each call, the team sees:

• The audio recording, right there — no need to go hunting in Dialpad.
• A word-for-word transcript. (It's about 80% accurate, which is exactly why having the recording alongside it is so handy.)
• The AI's best guess at which matter and contact the call relates to — plus a one-click option to add a new caller as a prospective client.
• An AI summary of the call.
• Suggested next actions, each with a suggested owner and due date.
• Suggested billables — how much, and to which matter.

An administrator or attorney reviews it all, cancels anything that's off, adds anything that's missing, and clicks Approve. Instantly, tasks land on the right teammates' dashboards and the billing gets noted.

That's only possible when everything flows into one secure place with your own private AI crunching it.

3. Matter pages built for your practice

Off-the-shelf practice management software tries to serve every practice area at once. Judith's experience with Clio: cluttered matter dashboards full of fields the firm never used, blank spots where nobody was sure what belonged, and information landing in the wrong place. Attorneys couldn't find things. Staff weren't sure where to put them.

And logging time? There were 17 different ways to do it. Seventeen. Easy to forget a timer running somewhere, then lose track of which matter it belonged to.

So when I asked Judith, "If you could have it any way you wanted, what would it look like?" — she designed it. Each matter type now gets its own custom layout: a progress checklist, related contacts, outstanding tasks, billables to date, and a timeline of every call and completed task. Anyone on the team can see where a matter stands at a glance.

And there's exactly one way to log time. It's always in the top-right corner.

4. "Where did we leave off?"

When I asked Judith what she wanted most, she didn't hesitate: "I want someone to just tell me where we left off."

This was the biggest thorn in the firm's side. DCZ Legal is a team of seven, and nearly everyone touches every case. Step away from a matter for even a week, and you'd have to check the latest pleadings in Google Drive, the notes in Clio, and the most recent emails in Gmail. Even then, you'd usually still have to ask someone. Hence all the team meetings.

Now there's an Ask Anything box on Judith's home screen. She types, "Where did we leave off on Alvarez?" Seconds later, there's the summary, with a link to the matter. She can follow up with "Who's the related party on Alvarez?" and get the answer, along with its source.

I'll sound like a broken record here, but it's true: that's not possible when your data is spread across five, seven, or twelve platforms. It's only possible when everything lives in one place and a private AI can comb through it for you.

So what does this cost?

This is one of the massive breakthroughs of AI.

A build like the one we just walked through could easily cost $500,000 to $1,000,000 from a software development company. You'd be looking at roughly six months of part-time work from computer scientists, developers, and user interface and user experience experts. For most law firms — even mid-size ones — that's been completely cost-prohibitive.

AI changed that.

Most of us think of tools like ChatGPT, Claude, or Gemini as helpers for writing emails or renaming files. But Claude has something deeper called Claude Code. Instead of helping you write a blog post, it helps you write software.

That might sound above your pay grade. I get it. I'm not a computer scientist. I don't know JavaScript. Yet I opened Claude Code and said, "I want to build a firm headquarters for Judith from scratch. Build it for me." It asked me questions. We went from there.

To be candid, it took time — roughly 250 hours of my time, plus Judith generously guiding me on what she wanted. But the Claude Code subscription over those few months didn't cost $500,000. It didn't cost $1,000,000.

It cost $400.

Think about that.

One important distinction: we use Claude to build the software. We don't use Claude to handle client data. The software we built — which we own — is what holds the data and runs the private AI.

Once it's running, enterprise-grade hosting (database and so on) runs about $2,000 a month. Here's the kicker: it's not uncommon for a firm to spend $15,000 a year on Clio alone. That one tool, on its own, nearly covers the cost of the entire headquarters.

And Clio is just the start. Add up the messaging, project management, and task management tools you no longer need — plus the labor you no longer have to pay for — and the savings across your whole tech stack go far beyond that.

And because you own it, you can customize to your heart's content. If Judith came to me every week for 52 weeks and asked, "Can we add a button here? Can we notify people this way? You know that deal where I landed 15 clients at once — can each of them log in and see the status of their case?" — the answer is yes. What else would you like?

It's like a genie lamp.

Where to start, based on where you stand

Judith sees three kinds of attorneys when it comes to AI:

• Dipping a toe in. Hesitant. Maybe using ChatGPT or Claude privately, but rightly worried about privacy and ethics.
• Wading in waist-deep. Using AI to draft emails, and using the AI features already built into tools like Lexis and Clio.
• Diving in headfirst. Like us. AI isn't just augmenting their work — it's doing a significant amount of the work.

If you're dipping a toe in: start with a time study

You don't have to do any of the high-tech things above. Start small, and don't move faster than you're comfortable with.

This is going to sound decidedly low-tech, but when Judith and I look back, our whole journey more or less started with a time study. A time study lays bare where your time is actually going. Maybe the fix is a tech tool. Maybe it's delegating something to an administrator. Maybe it's building a simple template with a friend — just like Judith and Chris did with that Excel spreadsheet.

But you can't fix any of it until you know where your time is going.

Grab ours for free at lawfirmpartners.com/time-study. You track your time for three days in a specially laid-out spreadsheet, and we show you how to spot the tasks that could come off your plate ASAP. Even if you never touch AI, it'll help you get time back for the work you got into law to do.

Once you've finished, strip out anything privileged or confidential (write "answered emails in Gmail," not the client's name and situation). Then paste this prompt into regular Claude or ChatGPT and attach your time study:

"Please look at my time study. What could I easily delegate to Claude or ChatGPT that doesn't require me to share client data and also gives me the opportunity to double-check my work before it goes out to make sure that it's accurate?"

I'll tell you right now: a lot of your time freedom is hiding in the boring stuff. Duplicating documents. Renaming files. Rinse-wash-repeat.

One more thing. If you use Claude or ChatGPT this way, turn off model training. In both tools, it's in your settings in the bottom-left corner — or just ask the chatbot, "How do I turn off model training?" and it'll walk you through it.

If you're waist-deep: take the time study to your team

You're probably not the only one losing hours to inefficiency. Have everyone on the team do the time study — attorneys, paralegals, clerks, and executive assistants.

At this stage, step up to a Claude Team plan. You can manage everyone in one account, create skills and share them with each other, and — critically — model training is turned off automatically.

Expect a question to come up, quietly or not so quietly: "Will AI take my job?" We wrote a whole article on exactly that, including the RIQ Reflection Exercise I've led with no fewer than nine teams. Every single time, team members walked out feeling empowered and excited. Read Will AI Take My Job?

If you're diving in headfirst: build your firm headquarters

Build in messaging, matter management, project management, a CRM, and whatever customizations your firm needs — call logging, mail processing, you name it.

You'll still keep a few third-party tools. We still use Dialpad; it simply sends its data into the headquarters automatically through an API. The attorneys still subscribe to Lexis and Westlaw. We still have Clio — we just use it far, far less, mostly for billing, contracts, and some calendar management. And we're not building our own email system. Google Workspace handles that.

Pro tips from the front lines

Mindset: hold a hackathon

In Judith's words, this is a grand adventure into uncharted territory. Dive in — with guardrails.

You'll make far less progress with an hour here and an hour there. Steal a page from the coding world and hold a hackathon day. I'd much rather spend five hours in one sitting than one hour on each of five days. There's momentum. And there's parallel processing: while AI builds one thing, you start a second, then a third.

When you look at your time study, look for tasks that are:

• Simple to teach a human — they'll be simple to teach AI, too.
• Repeated often — so the effort pays off.
• Low-risk if something goes wrong.
• Already on your plate — so you actually get the time back.

Skills: process first, code second

To build what we built, you'll need Claude Code, an understanding of process mapping (what happens, then what happens, then what happens), and the ability to work well with your team. Some web basics help, like buying a domain and pointing it to a host. If you don't have those, Claude can coach you through them while you build. A total beginner can get there. The more basics you have, the faster it goes.

Guardrails: protect the firm

• Get a BAA from every vendor that touches your headquarters. We use Supabase for our database, so we got a business associate agreement from them. Read the terms of even legal-specific tools, and you'll be shocked how few will sign one — meaning they aren't compliant for every kind of data in every situation.
• Address AI in your legal services agreement, and in your in-house policies for staff.
• Talk to a compliance or ethics attorney. I interviewed AI governance attorney Allison Mulford for our guide, Can Law Firms Use AI Tools Without Breaking Compliance?
• Keep a human in the loop. When an error isn't acceptable, let AI do most of the work, then have an administrator or attorney double-check before anything goes out — just like you'd proofread an email before hitting send.

And if you feel overwhelmed? I don't blame you. Start small and let it compound. Judith started with a time study, and look where we are today.

Who, not how: name an AI champion

Designate one person in the firm to lead the charge. If that's you, great. If not, maybe it's the tech-minded paralegal who's already the Clio whiz.

Why we're doing all this

We covered a lot of tech. Some of it may feel foreign. So let me close with what Judith told the room.

All of this technology exists to help you honor the reason you became an attorney in the first place. For Judith, that's helping people keep their agency, helping families protect what they've built and provide for the people they love, and helping every attorney at the firm be the best attorney they can be.

What's the reason you became an attorney? The law firm of the future gives you more time for exactly that.

We're cheering you on on this new frontier.

Here's to the future of your law firm,

Tim