The Legal Profession Has a 1 to 2 Year Window. Most Lawyers Will Miss It.

Summary of an essay from Dorna Moini's Legal Tech Notes, published on Substack on March 12, 2026. Dorna Moini is the CEO and Founder of Gavel and a former litigator at Sidley Austin. This page gives the argument and the numbers; the full essay is on Substack.

In brief: AI will not replace lawyers, but one lawyer with AI will replace several without it. The essay uses software engineering as the leading indicator: employment for developers aged 22 to 25 is down nearly 20% from its late-2022 peak, and the shift from skepticism to widespread AI use took about eighteen months. Law has the same structure and fewer places to hide, so firms have roughly one to two years to restructure around AI-assisted work before the leaders are set.

What happened to software engineers, and why does it matter for lawyers?

Engineering went first, and the pattern is visible in payroll data. A Stanford study of millions of payroll records found employment for developers aged 22 to 25 down nearly 20% from the late-2022 peak. Engineers who did not use AI were laid off first. Those who stayed moved up to architecture and system design, directing AI agents rather than writing code line by line, and because teams could now build more software, demand for senior engineering judgment grew.

Law is structurally similar. Statutes, regulations and case law are conditional logic written in natural language, which is why the same tools that write code can draft and analyze legal documents. The difference is what happens after automation.

Why does law lack the escape hatch engineering had?

Engineering escaped because automation created more products to build. Automating legal drafting does not create more laws to write. The essay also argues the "judgment" layer lawyers assume is safe is thinner than it looks: work that was too expensive for many clients can now be handled by fewer lawyers with technology, and that is where the volume goes.

The billable hour makes it worse. In engineering, AI made companies more profitable. In law, AI undercuts the revenue model itself. Personal injury firms can take a $15,000 supermarket slip-and-fall or a $20,000 wrongful termination claim that was uneconomical before; higher case volume then pushes courts and defense counsel to adopt AI too. New roles appear and disappear fast: "AI prompt engineer" was a hot title in 2023 and unnecessary by 2025, and e-discovery review teams are being automated. The essay predicts consolidation, with AI-capable firms taking most of the market.

What does the best-case firm actually do differently?

One early Gavel Exec customer, a firm representing startups in venture financings, loaded every Series B agreement it had ever negotiated against one venture firm (hundreds of agreements) into the system. In each new negotiation it asks the AI to redline the draft toward the most favorable terms that investor has previously accepted, citing the specific past deals. Every deal improves the system, the lawyers handle more deals, and competitors without that data cannot match the position. That compounding advantage, built on the firm's own negotiation history, is the essay's model of what to build now.

What should a firm do in the next year?

The recommendation is to restructure the firm around AI-assisted work rather than to "learn AI" or pay for the best model. Concretely: build AI-assisted workflows for the matter types you handle most, demonstrate them to clients, and treat the firm's own data, especially past negotiations, as the source of leverage. The essay's closing argument is that once the leaders in a practice area are established, some lawyers will not recover the ground.

Key numbers and sources from the essay

• Employment for software developers aged 22 to 25: down nearly 20% from the late-2022 peak (Stanford study of payroll records).
• About eighteen months from skepticism to widespread AI use among engineers.
• Cases that become economical with AI: a $15,000 slip-and-fall, a $20,000 wrongful termination claim.
• "AI prompt engineer": in demand in 2023, unnecessary by 2025.
• One Gavel Exec customer loaded hundreds of past Series B agreements negotiated against a single venture firm to drive redlines on new deals.
• Gavel is used by over 2,000 legal organizations in 28 countries.

What this means for a law firm or legal team

1. Pick the matter types you handle most and build an AI-assisted workflow for each, rather than buying a general tool and hoping adoption follows.
2. Treat your firm's own documents and negotiation history as an asset. A playbook built from your past deals is something no competitor can buy.
3. Show clients the workflow. Firms that demonstrate AI-assisted work win the clients who are already shopping on price and speed.
4. For contract work, Gavel Exec runs the playbook-based review described in the essay inside Microsoft Word; for repeatable documents, Gavel Workflows turns intake into finished drafts. Firms that have done this are in the Gavel case studies.

Frequently asked questions

Will AI replace lawyers?

The essay's answer is no, with a condition: lawyers who use AI will replace lawyers who do not. The evidence is software engineering, where engineers who did not adopt AI were laid off first and those who did moved into higher-judgment roles.

Why is the window one to two years?

Because that is roughly how long the same shift took in engineering, about eighteen months from skepticism to widespread use, and because once a few firms in a practice area have compounding data advantages, the rest cannot catch up by buying the same software.

How does AI change the billable hour?

AI makes small matters economical, so volume rises while hours per matter fall. Firms billing by the hour lose revenue per matter unless they move to flat fees and take on the volume AI makes possible.

Read the full essay: The Legal Profession Has a 1 to 2 Year Window. Most Lawyers Will Miss It. on Substack, or subscribe to Dorna Moini's Legal Tech Notes.

Related on gavel.io: Gavel Exec, Gavel Workflows, Gavel case studies.