You’re staring at the same spreadsheet every managing partner opens in January: another year of growth, another headcount decision. Your intake volume is up 30%, your associates are drowning in first-pass document review, and someone needs to answer the phone when a high-value prospect calls at 6pm on a Thursday.

The default answer is hiring. Another legal assistant at $55K base plus benefits. Maybe a junior paralegal at $65K if you can find one who won’t leave in eight months. You run the math, add 25% for payroll burden, and the line item lands somewhere between $70K and $85K per body.

But here’s the part that doesn’t show up in the budget: that hire solves exactly one slice of the problem. They work 40 hours a week. They take PTO. They can’t be in two places at once, and they sure as hell can’t review 400 pages of discovery documents in an afternoon without billing you $30 an hour for the privilege of moving slowly.

I’m not saying don’t hire people. I’m saying the cost comparison most firms run is incomplete. You’re weighing salary against salary, when you should be weighing capacity against capacity. Let’s walk through what that actually looks like.

The Real Cost of a Legal Assistant Hire

Start with the number everyone quotes: $55K to $75K for a legal assistant, $65K to $90K for a paralegal with three years of experience. Add another 20-25% for payroll taxes, health insurance, and the desk they’ll sit at. You’re at $70K to $110K all-in for the first year.

Now add the things nobody puts in the job posting. Two weeks to recruit. Another two weeks to onboard. Four to six weeks before they’re productive enough to take real work off your plate. During that ramp, someone senior is spending five to eight hours a week training them, which means five to eight billable hours you’re not invoicing.

If they’re good, they stay two years. If they’re great, a bigger firm poaches them at year three with a 20% raise you can’t match. If they’re mediocre, you spend six months managing them out and start the cycle again.

The hourly cost is easy to calculate: $70K divided by 2,080 hours is $33.65 per hour. But the effective cost is higher because you’re paying for lunch breaks, slow days, and the hours they spend learning your document management system instead of doing the work you hired them for.

And here’s the kicker: they can only do one thing at a time. When they’re on the phone with a client, they’re not reviewing intake forms. When they’re formatting a brief, they’re not answering the prospect who just submitted a contact form. You’re not buying 40 hours of work. You’re buying 40 hours of sequential task execution.

What Automation Actually Costs

An AI agent doesn’t cost $70K. For most firms in the $2M to $15M revenue range, you’re looking at $1,200 to $3,500 per month depending on the scope of work and the number of agents you deploy. Let’s call it $2,000 a month, or $24K annually, to keep the math simple.

That $24K buys you something fundamentally different than a human hire. It buys you parallel execution. An Intake Voice Agent can take six calls at the same time. A Matter Triage Agent can process 40 form submissions in the time it takes your assistant to read one. A Document Review Agent can work through a 300-page contract overnight and hand you a summary memo by 7am.

You’re not paying for hours. You’re paying for throughput.

The setup cost is real but contained. Most firms are live within three to four weeks. The first week is discovery: we map your intake flow, your matter types, your conflict-check process, and the handoff points where work currently falls through the cracks. Week two is build. Week three is testing with real data. Week four is go-live with your team watching the agent work alongside them.

There’s no ramp time in the traditional sense. The agent doesn’t need to learn where you keep the template library. It doesn’t forget how to run a conflict check. It doesn’t call in sick the day before a filing deadline.

And it doesn’t leave. You’re not replacing this capability in 18 months. You’re not conducting exit interviews or reposting the job on Indeed.

The Work You’re Actually Paying For

Let’s get specific about what these dollars are buying. Take client intake, the single highest-value workflow most firms underfund.

A legal assistant handling intake works something like this: phone rings, they pick up if they’re available. If the caller is serious, they take notes, ask qualifying questions, check for conflicts, and schedule a consult. If the assistant is good, this takes 12 to 18 minutes. If the call comes in after 5pm or during lunch, it goes to voicemail. Industry numbers suggest 30-40% of after-hours inquiries never convert because they’ve called two other firms by the time you call back.

Now compare that to an Intake Voice Agent. Every call is answered in two rings. The agent asks the same qualifying questions your best assistant would ask, in the same order, with the same tone. It runs a conflict check in real time against your matter database. It offers available consult slots and books the appointment directly into the partner’s calendar. The whole interaction takes eight minutes, and the agent sends a summary to your intake coordinator before the caller hangs up.

The math is straightforward: your assistant can handle maybe 15 to 20 intake calls a day if that’s all they do. The voice agent can handle 200. Your assistant costs $33 per hour. The voice agent costs about $4 per conversation when you divide monthly cost by call volume.

For document review, the gap is even wider. A junior associate billing at $250 an hour can review 20 to 30 pages of contract language per hour if they’re moving carefully. A Document Review Agent can process 200 pages in the same time, flagging every indemnity clause, every limitation of liability, every unusual termination provision, and producing a structured memo that reads like an associate wrote it.

You’re still going to have a senior attorney review the output. But you’ve collapsed four hours of associate time into 30 minutes of partner review. The client gets billed for 30 minutes instead of four hours, and your associate is freed up to work on something that actually requires legal judgment.

The Capacity Math That Changes Everything

Here’s where the comparison stops being about cost and starts being about what your firm can actually do.

Hiring a legal assistant gives you 40 hours a week of additional capacity. Hiring two gives you 80 hours. If you want to double your intake capacity, you need to double your headcount. The cost scales linearly, and so does the management overhead.

Automation scales differently. Adding a second agent doesn’t double your cost. It adds $800 to $1,500 a month depending on the complexity. Adding a third costs even less because the infrastructure is already in place. You’re paying for compute and API calls, not salary and benefits.

One firm we work with, a five-partner practice doing $8M a year, ran the numbers like this: they were spending $140K annually on two legal assistants who handled intake, matter admin, and first-pass document review. Good people, solid work, but they were the bottleneck. Intake calls during business hours were fine. After-hours calls sat until morning. Document review piled up whenever trial prep kicked in.

They deployed three agents: Intake Voice Agent, Matter Triage Agent, and Document Review Agent. Total cost: $32K annually. The assistants didn’t get fired. They got promoted. One moved into paralegal work on complex litigation matters. The other took over client communications and court filings, work that actually required human judgment and relationship management.

Intake conversion went up 22% because every call got answered. Document review turnaround dropped from three days to same-day. The partners stopped working weekends to catch up on admin. The firm added $1.2M in revenue the next year without adding headcount.

That’s not a typical result, it’s one data point. But the pattern holds: automation doesn’t replace your team, it changes what your team is capable of doing with the same number of hours in the day.

What You’re Really Buying With Either Option

When you hire a legal assistant, you’re buying reliability within a narrow band. They show up, they do the work, they go home. If they’re good, they get a little faster over time. If they’re great, they start to anticipate what you need before you ask. But they can’t be in two places at once, and they can’t work faster than human speed.

When you deploy an AI agent, you’re buying leverage. The agent works at machine speed on repetitive tasks, which frees your human team to work on the tasks that actually require empathy, judgment, and client trust. You’re not choosing between humans and machines. You’re choosing between a team that’s constantly underwater and a team that has room to think.

The cost difference is real: $70K to $110K per assistant per year versus $24K to $42K for a multi-agent setup that handles three times the volume. But the capacity difference is the part that changes your business. You’re not just saving money. You’re removing the ceiling on how much work you can take on without hiring your way into a bloated cost structure.

If you want to see what this looks like for your firm specifically, we built a process for that. It’s called an Omni Audit, and it’s the opposite of a software demo. Sixty minutes, three outputs: a workflow map of where your team is spending time on repetitive work, a capacity model that shows you what automation would actually unlock, and a build roadmap with cost and timeline for the agents that make sense for your practice. No deck, no sales pitch, just the numbers your next partner meeting needs. Book a 60-min Omni Audit and we’ll walk through it together.

The Workflow Breakdown: Where Automation Wins

Let’s take three workflows most firms run manually and compare what each option actually delivers.

Client intake and qualification. Your legal assistant picks up the phone, asks questions, takes notes, checks conflicts manually or via your practice management system, and schedules a consult. Total time: 15 minutes per call. Calls handled per day: 18 to 22 if intake is their only job. After-hours calls: voicemail.

An Intake Voice Agent picks up every call in two rings, 24/7. It asks the same questions in the same order, pulls conflict data in real time, books the consult, and logs everything in your CRM before the call ends. Total time: 8 minutes per call. Calls handled per day: unlimited. After-hours calls: answered immediately.

Cost per interaction: $33/hour for the assistant divided by four calls per hour is $8.25 per call. The voice agent runs about $4 per call at typical volume. You’re paying half as much and capturing 100% of inquiries instead of 60%.

Matter triage and routing. Your assistant reviews intake forms, reads the case summary, decides which partner should see it, and forwards the email with a note. Time per matter: 10 to 15 minutes. Matters processed per day: 25 to 30.

A Matter Triage Agent reads the form, classifies the practice area using the same criteria your assistant would apply, scores the fit based on case value and complexity, and routes it to the right partner with a one-paragraph brief attached. Time per matter: 90 seconds. Matters processed per day: 400-plus.

The assistant costs $4.50 per matter at 30 matters a day. The agent costs about $0.40 per matter. You’re paying a tenth as much and processing ten times the volume.

First-pass document review. A junior associate reads a 200-page contract, flags key clauses, checks for standard provisions, and writes a memo summarising risk and recommending revisions. Time: 6 to 8 hours. Cost at $250/hour: $1,500 to $2,000.

A Document Review Agent reads the same contract in 45 minutes, flags every clause type you care about, cross-references your standard positions, and produces a memo in the same format your associates use. A senior attorney reviews the output in 30 minutes and bills the client for one hour. Total cost: $250 to the client, $12 in agent compute cost to you.

The client pays less. You make more margin. The associate works on something that actually requires a law degree.

You can see the full breakdown of what an AI audit looks like for law firms at the AI audit for law firms page. It walks through the discovery process, the workflow mapping, and the three outputs you’ll walk away with.

The Hidden Costs Nobody Talks About

Hiring has friction costs that don’t show up in the salary line. Recruitment takes time. Onboarding takes time. Training takes time. Management takes time. Every one of those hours is a partner or senior associate not billing clients.

If you spend 20 hours recruiting, 15 hours onboarding, and 30 hours training over the first three months, that’s 65 hours of senior time. At a blended rate of $300/hour, that’s $19,500 in opportunity cost before the new hire is productive. Add that to the $70K salary and you’re at $89,500 in year-one cost.

Automation has setup cost, but it’s contained and predictable. Discovery and build take three weeks. Your team spends maybe 10 hours total in working sessions. Once the agent is live, management overhead is close to zero. You’re not conducting performance reviews. You’re not mediating disputes about PTO. You’re not worrying about whether they’ll leave for a competitor.

The other hidden cost is error rate. A legal assistant working fast will miss things. They’ll route a matter to the wrong partner. They’ll forget to run a conflict check. They’ll schedule a consult in a slot that’s already blocked. It happens. They’re human.

An AI agent doesn’t get tired. It doesn’t skip steps. It runs the same process the same way every time. If you tell it to check conflicts before booking, it checks conflicts before booking. The error rate on repetitive tasks is functionally zero.

That consistency has a dollar value. One missed conflict check can cost you a client relationship or worse. One missed intake call can cost you a $50K case. The cost of those errors is hard to quantify until they happen, but they’re real.

What the Next 18 Months Look Like

If you hire a legal assistant today, here’s what the next 18 months look like: Month one is recruiting and interviewing. Month two is onboarding. Month three is when they start to pull their weight. Months four through twelve, they’re productive and things are better than they were. Month thirteen, they get a job offer from a bigger firm. Month fourteen, you’re back to recruiting.

If you deploy an AI agent today, here’s what the next 18 months look like: Week one is discovery. Week two is build. Week three is testing. Week four, you’re live. Months two through eighteen, the agent is handling the same volume it handled in month one, except now your team has built workflows on top of it that didn’t exist before. The agent doesn’t leave. It doesn’t get poached. It doesn’t ask for a raise.

The compounding effect is what matters. A human hire gives you linear capacity growth. You add one person, you get one person’s worth of work. An AI agent gives you exponential capacity growth because it frees your existing team to do higher-value work, which generates more revenue, which funds more automation, which frees more time.

Firms that start this process now are 18 months ahead of firms that wait. They’ve already rebuilt intake. They’ve already automated document review. They’ve already redeployed their assistants into client-facing roles that actually require human judgment. They’re not wondering whether automation works. They’re wondering what to automate next.

The Practical Path Forward

You don’t have to rip out your entire operation and replace it with robots. Most firms start with one workflow. Intake is the most common because it’s high-value and the ROI is immediate. You deploy an Intake Voice Agent, you run it in parallel with your existing process for two weeks, and you watch what happens.

If it works, you expand. You add a Matter Triage Agent to handle form submissions. You add a Document Review Agent for contracts. You keep your legal assistants and paralegals, but you change what they spend their time on. The repetitive work goes to the agent. The judgment work stays with the human.

The firms that do this well don’t treat it like a technology project. They treat it like a capacity project. The question isn’t “Can we afford automation?” The question is “Can we afford to keep running the same process that caps our growth at 40 hours per human per week?”

If you want a practical checklist for how to think through your intake process specifically, we built a worksheet that walks through the decision points. It’s called the AI Client Intake Checklist for Law Firms, and it covers the questions you need to answer before you deploy anything: what data you need, how to handle conflicts, how to route calls, and how to measure whether it’s working. You can download it here: AI Client Intake Checklist for Law Firms. It’s a working document, not a marketing brochure.

The ROI comparison is simple when you put the numbers side by side. A legal assistant costs $70K to $110K per year and gives you 40 hours a week of sequential task execution. An AI agent setup costs $24K to $42K per year and gives you unlimited parallel execution on the tasks that don’t require a law degree.

You’re not choosing between people and technology. You’re choosing between a team that’s constantly at capacity and a team that has room to grow. The firms that figure this out in 2025 will be the firms that don’t need to have this conversation in 2027 because they’ve already built the leverage.

If you want to see what this looks like for your firm, the next step is an Omni Audit. Sixty minutes, three outputs, no deck. We map your workflows, model your capacity, and show you the build roadmap with real costs and timelines. Book my Omni Audit and we’ll walk through it together. You’ll know exactly what automation would cost, what it would unlock, and whether it makes sense for your practice before you spend a dollar.

The cost comparison is just the starting point. The capacity comparison is what changes your business. Most firms are still running the math on salary versus salary. The firms that win are running the math on what their team can actually accomplish when the repetitive work is handled by something that never sleeps, never leaves, and never caps out at 40 hours a week.