The State of Law Firm Growth in 2026: What 6 Experts Told Us

Read with : ChatGPT | Claude | AI Mode

A client calls your office already holding a ChatGPT summary of their legal problem. Your ads are bringing in leads, but half of them go cold before anyone calls back. Revenue was up this year, but somehow there’s less left over at the end of the month than there should be. You know you should probably be doing something with AI, but you’re not sure if using it puts client data at risk.

None of that is hypothetical. Those are the exact problems six different experts addressed, one at a time, in a series of webinars hosted by Law Firm Growth Ideas over the past several months. Each session covered one specific part of running a firm: keeping AI use secure, using AI to handle client intake, fixing how the firm actually runs day to day, what happens when clients research their case with AI before ever calling you, the financial mistakes that quietly eat into profit, and how to get known for your work without living at networking events.

Below is what each expert said, one section per webinar, with the source credited and linked. Read it start to finish, or jump straight to whichever section matches the problem you’re actually dealing with this week.

AI adoption isn’t the risk; poor implementation is
Intake is where AI actually pays off
Your firm doesn’t have a growth problem; it has an operating problem
Clients are already asking AI about their case
Revenue isn’t the same as profit
You are the brand, not your firm name
The pattern across all six conversations
Common mistakes firms make right now
Frequently asked questions
Where to start this quarter

1. AI adoption isn’t the risk; poor implementation is

The security risk everyone worries about isn’t really the AI tool. It’s what happens before anyone ever opens it. Matthew Kaing (Founder, eSudo) and Sharon Means, CPA, said it directly: “The risk does not come from AI, but from poor implementation, unclear rules, and missing oversight.”

Kaing’s advice comes down to a phrase he used more than once: treat AI like a super intern. A good intern does the fast research and gives you a draft. A good intern does not get to make the decision or send the proposal without you checking it first. That’s the whole model. Sort your data into public, internal, and confidential, with “confidential” meaning client files and anything financial or medical—the kind of information that should never touch an unsecured tool. Only use products that don’t train on your data and let you actually control permissions. Write down what’s allowed. Check the settings regularly, because a well-known tool that’s misconfigured creates more risk than an obscure one that’s locked down correctly, and Microsoft Copilot itself behaves differently running in a private Azure environment versus routing through public servers.

Two specific things worth knowing that don’t show up in a generic AI policy. First, Kaing’s own rule on AI browsers: “No AI browser use for work is the policy we have right now,” because they read everything visible on screen, including whatever client file happens to be open in the next tab, even without giving the tool access to your inbox. Second, when someone on the call asked whether attorneys should disclose AI use in their client engagement agreements, Kaing’s honest answer was that he’d seen people start putting it in their agreements as a simple disclosure, and as more tools bake AI in by default, it’s probably heading toward standard practice rather than a red flag.

Full breakdown: The 6-Step Safe Path to AI Adoption for Law Firms

2. Intake is where AI actually pays off

Your marketing budget usually isn’t the problem when campaigns underperform. Intake is. Danny Abir, managing partner at ACTS LAW, a firm that’s recovered over a billion dollars for clients since 2016, put a number on why speed matters: a firm that responds slowly to a new lead is paying to generate a lead it then loses, regardless of how good the campaign was.

The clearest example he gave: a batch of 1,550 individual client submissions for the firm’s LA County camp abuse settlement, the kind of work that normally takes about two months, got done in one week using an AI tool the firm built internally, with attorneys still reviewing every submission before it went out. That’s the model across every use case Abir described. AI drafts, a person checks, and nothing client-facing skips that step.

His sharpest line on where this is heading, worth sitting with regardless of your practice area: “AI is not going to replace lawyers, but AI, those attorneys who know how to use AI will replace the ones that don’t know how to use AI.” He also raised something specific to firms that bill hourly: if AI makes research and drafting faster, does that mean less billable time, and does the billable hour model itself start to break down? His read is that this eventually pushes firms toward value-based pricing rather than hourly billing, the same way legal research tools already compressed hours without anyone renegotiating rates line by line.

One mistake worth flagging directly from his answer to an audience question: putting client information into free AI tools risks waiving attorney-client privilege, because that data isn’t protected the way information shared with your attorney is.

Full breakdown: How Law Firms Are Using AI to Win More Clients

3. Your firm doesn’t have a growth problem; it has an operating problem

If your firm feels busy but out of control, the fix is probably not a new system. It’s actually using the systems you already paid for. Janelle Sam, who holds an MBA from Cornell, told a story from her work at Sebrook Law Offices that makes this concrete. Clients kept complaining their bills were too high. Staff insisted the clients just didn’t want to pay. Sam pulled the firm’s case management data, Cleo, and found the actual cause: every time a client sent an email, everyone copied on it the attorney, the paralegal, and the legal assistant were separately logging billable time just for having received it. Nobody saw it as a problem because nobody was looking at the data across every invoice at once, only one bill at a time. Once the firm regrouped its billing categories and fixed the policy on what gets charged, client complaints and negative reviews dropped almost overnight.

Her point wasn’t really about Cleo. It’s that most firms are running expensive software far below its capacity, using it to generate invoices and nothing else, because attorneys went to law school, not business school, and the tool’s real reporting power never gets touched.

Her broader diagnosis, told through a fire metaphor: “It’s like seeing a building burning. You see the flames over there, but guess what? The gas needs to be turned off… instead you’re beating the flames, beating the flames, and the fuel is pumping into it. ” A firm chasing more leads because conversion feels weak is usually beating the flames. The actual leak is often somewhere else, in pricing that never adjusted as the firm added staff, or in a process nobody owns.

Her sharpest line on scaling: “You cannot outgrow your systems.” A firm that grows past what its current systems and people can handle doesn’t get more successful; it breaks, the same way a bicycle built for one person doesn’t work with five people on it.

Full breakdown: The Freedom Framework: Reclaiming Time, Profit, and Control

4. Clients are already asking AI about their case

A client who shows up having already asked ChatGPT about their case isn’t a threat to your practice. The gap between what AI just gave them and what you can give them is your best argument for why they need you. Ben Barry (COO, Moradi Neufer LLP) and Joey Tran (Partner, Acceleron Law Group) put the whole relationship in one line: “These AI agents, they give data, they don’t give wisdom or expertise… they should be used as tools, not as advisors.”

Barry gave a concrete version of where this goes wrong: clients now show up asking things like “Why aren’t we doing an RFO for this?” using technical terms they picked up from a chatbot without understanding what the term actually means or whether it applies to their situation. He also described a pattern worth knowing before your next consultation: younger clients tend to bring more AI-generated material into the conversation but are also more willing to be corrected. Older clients who invest time in an AI answer tend to hold onto it harder, even when it’s wrong, because it feels like it came from research rather than an opinion.

Tran ran his own test that’s worth understanding directly. He fed a landlord-tenant contract dispute into an AI chatbot and got a strong argument for his position. Then, just to see what would happen, he asked it to argue the opposite side using slightly different framing, and it built an equally convincing counterargument. His conclusion wasn’t that the tool failed. It’s that if he’s using AI to strengthen his case, opposing counsel is doing the same thing, and the actual outcome still comes down to which attorney knows the judge, the court, and how to defend the argument in the room, something no chatbot has access to. He made the same point about contract language specifically: an AI tool can suggest wording, but it can’t defend to opposing counsel why the contract says “shall” in one place and “may” in another, and a practitioner still has to be able to answer that question word by word.

Barry’s example on drafting: prenups. AI can produce a reasonable starting document. What it can’t do is walk a couple through the uncomfortable conversations about money and kids that a prenup process is actually supposed to surface, the part that makes the document worth having in the first place.

One more detail worth checking for yourself: Tran said he now asks AI tools for a source link on anything factual, specifically because “sometimes the link that I go to it no longer exists” or sits behind a paywall he can’t verify without paying for it separately.

Full breakdown: 12 Things AI Can’t Give Your Clients

5. Revenue is not the same as profit

More revenue can mean less profit, and most firms don’t find out until it’s already happened. Rush Shah, a fractional CFO based in Los Angeles, opened his session with the line that frames everything else: “Growth without margins is just a losing strategy in my opinion.”

His first fix is a mindset one: stop tracking actual-versus-budget, a backward-looking snapshot of where you stand today, and start tracking forecast-versus-target, which tells you whether you’re actually going to hit your number or not. He also drew a distinction worth checking your own hire against: not every CFO does the same job. A startup CFO is focused on fundraising and burn rate. A service-business CFO, the kind a law firm needs, is focused on utilization, project timing, and cash flow tied to matters. Hiring the wrong type means paying for advice that doesn’t fit how your firm actually makes money.

The clearest number from the whole session: a trust and estates firm handling 379 matters a year at about $5,000 each brought in $1.895 million in revenue at a 25 percent margin. To hit a 35 percent margin, Shah worked backward and found the firm needed roughly 58 additional matters at a slightly higher price, around $5,500 average, or a minimum hourly rate near $367. Here’s the part worth sitting with: if that same firm had instead just chased the same higher profit number through revenue alone, keeping its old pricing the same, it would have needed somewhere between 200 and 300 new matters to get there, five times the workload for the same result. Same firm, same staff capacity, wildly different amount of work required, depending on whether pricing or revenue leads the plan.

Full breakdown: The 3 Financial Mistakes Quietly Killing Law Firm Profit

6. You are the brand, not your firm name

By the time a prospective client calls your office, they’ve usually already decided whether they trust you. That decision gets made somewhere else entirely, before the phone even rings. Steve Fretzin, who has coached legal business development for over 18 years, put it plainly: “You are the brand” in this profession, because clients hire a person, not a firm name.

His example of what a real, working differentiator looks like: the Chicago employment law firm Laner Muchin found through client surveys that their clients cared more about getting a call back fast than almost anything else, so they built a two-hour return-call guarantee and said so publicly. Nobody else in their market was making that specific promise, and it mattered more to their actual clients than another claim about expertise would have.

On what to actually publish: Fretzin’s own experience is that lawyers gravitate toward podcasts over video, specifically because they can listen while doing something else, driving, walking the dog, or commuting, in a way a video demands full attention they don’t have. He was also direct about a trap worth naming here: “You don’t want to do AI slop because you can get tagged for that now on LinkedIn,” and his fix is blunt: get rid of anything that reads like generic AI output and keep it in your actual voice, a point that lands harder given everything above about clients already comparing your content against a chatbot’s answers.

His repurposing system in practice: he records podcast interviews, runs the transcripts through AI to generate first-draft articles, and sorts those into categories, and he uses that same process to assemble an entire book with contributions credited to everyone he interviewed. One idea becomes five formats without five separate efforts. And if writing or recording isn’t realistic yet, his fallback is simpler than most people expect: comment consistently and usefully on the posts of the people you want to stay visible to. It costs less time than posting and still keeps your name in front of them.

Full breakdown: Cut Your Networking Time in Half: A Personal Branding Playbook

The pattern across all six conversations

Looked at one at a time, these are six unrelated topics. Looked at together, the same few ideas keep showing up from different directions.

Figure out the actual problem before buying anything. Kaing says map the workflow before buying an AI tool. Sam’s Cleo story is the same lesson from a different angle: the firm already owned the tool that would have solved its billing complaints; it just wasn’t being used correctly.

AI gives you information. A person gives you judgment. Barry and Tran made this point about AI versus lawyers directly. Shah made almost the same point about bookkeeping versus a CFO. Reports tell you what already happened. A person decides what to do next.

A person checks the work every time. Every session that touched on AI said the same thing without exception, and Abir’s mass tort example shows what that looks like at scale: AI produces the first pass, and an attorney reviews every single one; nothing client-facing skips that step.

Growing without a target doesn’t mean you’re actually growing. Shah’s 58-versus-300-matters comparison and Sam’s building-fire metaphor are both describing the same habit: chasing more without deciding in advance what “better” actually means.

Clients are forming opinions before they ever call you. Fretzin’s point about being researched online and Barry and Tran’s point about clients arriving with AI research describe the same shift, just from opposite sides of the phone call.

Common mistakes firms make right now

Buying an AI tool before figuring out what it’s supposed to fix. Both Kaing and Sam warned about this from different angles, and Sam’s Cleo story shows how much value sits unused in tools firms already own.

Assuming a well-known AI tool is automatically a safe one, or that AI browsers are harmless. Kaing’s own rule against AI browsers at work exists specifically because they read whatever else is open on screen.

Putting client information into free AI tools. Abir’s warning is direct: this risks waiving attorney-client privilege, since that data isn’t protected the way information shared with your attorney is.

Treating “profitable” and “in control” as the same thing. A firm can be making real money while nobody but the owner can actually run it, which is exactly what broke down in Sam’s Cleo example before anyone looked at the data.

Chasing more revenue instead of designing for margin first. Shah’s numbers make this concrete: the same profit goal took 58 additional matters with the right pricing, or 200 to 300 without it.

Publishing generic content, or obvious AI-generated content, instead of sounding like an actual person. Fretzin’s warning about AI slop getting flagged on LinkedIn is the sharpest version of this, and it matters even more now that clients are actively comparing firm content against what a chatbot told them.

Frequently asked questions

Should my firm start using AI now or wait until it’s more established?

Every expert pointed in the same direction: start now, but start with a policy and a clear idea of what data is off-limits. Abir’s point that AI-literate attorneys will replace AI-illiterate ones, not that AI replaces attorneys, is the clearest reason not to wait.

What’s the one thing I should fix first?

Based on these six conversations, intake, and specifically how fast you respond to a new lead. It’s the point where marketing spend either turns into revenue or gets wasted.

If my revenue is up, is my firm actually doing better?

Not necessarily. Shah’s own numbers show a firm can hit the same profit goal with a fraction of the added workload, just by pricing for margin first instead of chasing revenue and hoping margin follows.

Is AI going to replace the need for a lawyer?

Based on Barry and Tran’s own testing, no. AI can generate a strong argument and an equally strong argument for the opposite side in the same conversation. Judgment about which one actually wins in front of a specific judge is still a human job.

Where to start this quarter

Five things worth doing based on everything above.

Write down, in one page, what data your firm currently lets into AI tools, using Kaing’s public, internal, and confidential buckets, and fix anything confidential that shouldn’t be there.

Pull a report from whatever case management software you already pay for and look at what it’s actually tracking, the way Sam found the email-billing problem hiding in Cleo data nobody had looked at.

Time how long it actually takes your firm to respond to a new lead, and compare that against what you’re spending to generate it.

Run Shah’s math on your own numbers: what would it take to hit your next profit goal through pricing versus through pure volume, and which one is actually less work.

Write one piece of content this month that sounds like you, not like a chatbot, answering a real question a client has asked recently, and check whether it shows up when you ask an AI chatbot that same question.