AI Just Compressed Your Production Costs. Now What?

Imagine you’re looking at 9,000 pages of discovery documents. Verifying completion, documenting what’s there, synthesizing it all into something useful. Twenty hours of work, maybe more. Now someone tells you AI tools can compress that to five hours. Maybe less.

That’s good news, right?

Except those 15 hours were billable. Where does that revenue go?

This is the question I keep hearing when I talk to people in and around the legal industry. Not “will AI change legal work?” That part is settled. The question is what happens to the business model when the production costs collapse but the billing structure hasn’t caught up.

It’s a pattern I’ve seen before in other industries. When technology compresses the cost of production, the businesses that restructure around the new economics win. The ones that try to preserve the old model lose ground. Slowly at first, then all at once.

This isn’t about AI replacing lawyers. The judgment, the strategy, the counsel you provide. That’s still the core value. But AI is changing what it costs to produce the work that surrounds that judgment. And that changes everything about how you price it, staff it, and compete for clients.

What AI Is Actually Doing in Law Right Now

I want to start with what’s already happening, not what might happen someday. The tools are live. According to the Clio Legal Trends Report for 2025, 79% of legal professionals now use AI in some form.

What does “use AI” actually mean in practice? Document review and summarization. Legal research that pulls relevant case law in seconds instead of hours. First-draft generation of motions, contracts, custody agreements, and letters. Not final work product. But the most time-consuming part of producing it.

Here’s the stat that makes this concrete: according to research cited by the California State Bar’s own guidance, AI can automate up to 74% of the work that law firms currently bill hourly. Not replace lawyers. Automate the production work.

74% of the work law firms currently bill hourly can be automated with AI. If you’re a solo or small firm practitioner, the firms with 51+ attorneys are already acting on that number.

Take family law as an example, because it’s document-heavy and paralegal-intensive, which makes the illustration clear. Custody motions. Financial disclosure review. Discovery requests. All of this is exactly the type of work AI handles well. The first draft that used to take four hours can be generated in 20 minutes, then refined and reviewed by an attorney who brings the judgment the tool can’t.

But here’s the part that matters if you’re running a solo practice or a small firm: adoption is uneven. Firms with 51 or more attorneys use generative AI at roughly double the rate of solo and small firms. The big shops are already there. You’re competing against them for the same clients. The gap isn’t about whether the technology works. It’s about who moves first.

The Money Behind It

This is the part where the numbers matter, because the numbers are what makes this a business story and not just a technology one.

Legal tech companies raised $5.99 billion in 2025. Fourteen companies raised $100 million or more in a single year. Harvey AI hit an $11 billion valuation as of February 2026, up from $8 billion just one month prior. For context, that valuation is larger than most of the companies these tools are being built to serve.

Clio, the practice management platform that most small law firms already use, hit a $5 billion valuation in November 2025. Thomson Reuters launched agentic AI workflows for CoCounsel in early 2026: autonomous document review, multi-step legal analysis without human intervention. The legal AI market hit $7.2 billion in 2025 and is on track to exceed $10 billion by end of 2026.

When capital concentrates at this scale around a specific technology, it finds its way into the work. This money is building tools aimed at your practice area. The question is whether you adopt them or compete against firms that did.

California Gave You a Framework

Here’s the part that surprised me when I went looking. California has already thought about this. And the framework they built is actually better news for practitioners than most people realize.

The California State Bar published practical guidance on generative AI in November 2023, then expanded it in May 2025. These aren’t new laws. They’re the application of existing professional conduct rules to a new context. Six duties are explicitly covered: confidentiality, competence, supervision, billing, candor to courts, and anti-bias.

Start with confidentiality, because it’s the floor. The guidance draws a hard line on client data. Firms using consumer-grade tools (like the free tier of ChatGPT) with client information are already in violation. If you’re going to adopt AI, this is the minimum. And it’s a differentiator. The firms that get this right can tell clients they did.

Billing transparency is the one most practitioners feel in their gut. If AI saves time, you can’t ethically charge the same as if it didn’t. The guidance is explicit: bill for time actually spent, including refining prompts and reviewing AI output. Don’t bill for time the AI eliminated.

But here’s the reframe. That guidance isn’t a revenue cut. It’s a pricing redesign opportunity. The State Bar just gave you ethical cover to move to value-based pricing.

And they know it. Their own proposed formal opinion on flat fees (Interim No. 20-0003, published November 2025) explicitly acknowledges that flat fee use has “expanded beyond traditional settings” and is “increasingly adopted by large firms, driven by advanced legal technologies.” The Bar isn’t just permitting this shift. They’re watching it happen and building the framework around it.

Disclosure is required. Lawyers must tell clients when AI is involved. That sounds like a burden until you hear it as a pitch: “We use AI for research and first drafts, which lets us move faster and keep your costs down. Here’s how we supervise the output.” That’s not a compliance checkbox. That’s a client development conversation.

And supervision (Rule 5.3) means you remain responsible for all AI-assisted work. The lawyer can’t outsource ethical responsibility to a tool. That’s actually protective. It means the value of your judgment doesn’t go away. It just sits on top of a faster production stack.

The Day Anthropic Entered the Arena

On February 2, 2026, Anthropic (the company behind Claude, which powers Cowork, the tool I use every day) released a legal plugin as part of its agentic Cowork capability. Contract review, NDA triage, compliance workflows, legal briefings, templated responses for common inquiries. Not giving legal advice. Assisting the workflow.

The market reaction was immediate. Thomson Reuters stock dropped 16%. Wolters Kluwer dropped 10%. These are the companies that sell legal research and workflow tools to law firms. The market read this as a direct competitive threat.

Thomson Reuters dropped 16% in a day. Foundation model companies aren’t just building infrastructure anymore. They’re building toward the end product. That changes the math for every firm buying legal tech.

Above the Law’s take was blunt: “What Anthropic’s release of Claude Legal Skills means for solos and smalls: nothing.” Their argument was that Cowork requires enough technical setup that small firms won’t use it without help. And honestly, they’re right about that part. It won’t self-install into a solo practice.

But it signals something bigger. Foundation model companies are no longer just building infrastructure. They’re building toward the end product. That has implications for how long the specialized legal AI vendors can hold their pricing and their margins.

Here’s why that matters for your practice: when competitive pressure from that direction compresses vendor pricing, the tools you’ll eventually use are going to cost less than the early adopters paid. The window between “expensive” and “accessible” is closing faster than most people expect.

Three Business Model Decisions You’re Going to Face

This is the part where the technology story becomes a business story. If you run a firm, three decisions are coming at you. All three are opportunities if you move early enough.

1. Flat fee vs. hourly: capturing the value of your efficiency.

If AI can produce a first-draft custody motion in 20 minutes instead of four hours, the cost to produce that motion just collapsed. Hourly billing, in that environment, means watching revenue evaporate hour by hour. But flat-fee pricing lets you capture the value of your efficiency gains.

This isn’t a fringe idea. Fifty-nine percent of all firms now bill flat fees exclusively or alongside hourly, according to Clio’s 2024 data. Only 41% bill exclusively by the hour. Flat fee billing is no longer the alternative. It’s becoming the default. And the performance data backs it up: firms billing flat fee collect payment twice as fast and close matters 2.6 times faster than firms billing hourly.

Seventy-one percent of clients already say they’d prefer a flat fee over hourly billing. The firm that prices a flat-fee motion at $800 when the old hourly cost was $1,200, and produces it in a quarter of the time, just improved margins and won the client. That’s not a loss. That’s a business model upgrade.

2. Speed as a service advantage: winning clients, not just cases.

Law has always competed on speed. But it used to be a question of attorney availability and paralegal bandwidth. AI flips that. A firm with AI-assisted research and drafting can turn a motion in hours instead of days. The firm that can credibly say “we’ll have your motion filed by end of day” is winning the client before the first billable hour starts. That’s a different kind of competitive advantage, and it accrues to the firms that adopt early.

3. Transparency as competitive positioning.

The firms that explain their AI workflow are building trust that becomes a competitive moat. “We use AI for research and first drafts. Here’s how we supervise the output. Here’s how that keeps your costs down.” That’s a client development conversation, not a confession.

The California State Bar guidance makes disclosure mandatory. But beyond compliance, transparency is a relationship signal. The firms using AI quietly and billing as if nothing changed are one client conversation away from a problem. The firms that lead with it are one conversation away from a new client.

What an Operator Sees Looking In

I’m not a lawyer. I’m an operator who watches what happens when technology compresses the cost of production. I’ve seen this pattern before, in other industries, with other technologies. The specifics change every time. The shape of the disruption doesn’t.

When production costs collapse, the businesses that restructure around the new economics win. The ones that try to preserve the old billing model lose. Not immediately, but steadily. The clients notice before the firm does.

The firms that figure out where the freed hours go are the ones I’d want to work with. Not because they use the shiniest tools, but because they’ve thought about what the tools mean for their business and their clients. That kind of thinking is the actual differentiator.

This isn’t about AI. It’s about what happens when the cost of producing the work drops and the value of the judgment stays the same. The firms that price for the judgment, that use the freed capacity to serve more clients, move faster, and build deeper relationships. Those are the firms that come out ahead.

If you’re a solo or small firm trying to figure out where to start

I’ve been designing AI workflows for legal practices using general-purpose tools and existing practice management software, not specialized legal AI subscriptions. The architecture handles California State Bar compliance (confidentiality, billing transparency, supervision, disclosure) and leaves the attorney owning the system. No vendor lock-in, no ongoing dependency.

If you want to talk through what this looks like for your practice, I’m at bmceachen@gmail.com.

This is Earned in the Fire: hard-won lessons on operations, AI, and building what works. If this was useful, subscribe for future essays on what happens when technology reshapes how businesses run.