Listen to the article
A third-year associate is staring at a contract that she did not draft somewhere on the fourteenth floor of a Midtown Manhattan office. Software completed the initial pass. Yes, she is marking it up and adding the kind of judgment calls that cost $1,200 per hour. However, the lengthy, laborious process of reading each page from the beginning—what senior partners used to refer to as “earning your spurs”—has simply vanished. At dinner, she shared it with a friend. He chuckled. He works as a litigator for another firm and stated that the current discovery process is the same. In an afternoon, millions of emails were sorted.
This is the peculiar transitional phase that the legal sector has entered. Big businesses aren’t bragging about AI, and small businesses aren’t acknowledging how much they rely on it. Beneath the surface, in the silence between deadlines, the change is taking place. McKinsey’s estimate, which states that 22% of a lawyer’s work can already be automated and 44% could theoretically be, seems more like a description of last Tuesday than a forecast. It appears that investors also believe it. Harvey AI alone reached a $8 billion valuation in December, and the legal AI market doubled from $1.5 billion to over $3 billion in just one year.
| Industry | Legal services & legal technology |
| Market Size (2024) | $1.5 billion |
| Market Size (2025) | Over $3 billion |
| Projected Market Size (2030) | $10.8 billion |
| Tasks Currently Automatable | 22% of a lawyer’s daily work |
| Total Technically Automatable | 44% of all legal tasks |
| UK Lawyer Generative AI Use | 61% (up from 46% eight months prior) |
| U.S. Legal Professionals Using AI | Around 79% |
| Notable AI Legal Tool | Harvey AI — valued at $8 billion (Dec 2025) |
| Expected Reduction in Chargeable Hours | ~16% across UK top 100 firms |
| Key Concerns | Privacy, accuracy, hallucinations, shadow AI use |
| Primary Use Cases | Contract review, legal research, document drafting, e-discovery |
Speaking with individuals within companies, it seems like no one is quite sure how to handle this. Attorneys are typically cautious adopters. The idea that a sentence in a contract might mean something different in fifteen years, caution, and demonstrating your work were the foundation of the profession.
They are now employing citation-hallucinogenic tools. Akber Datoo, who owns a legal tech company in London, talked about having dinner with a former US CFTC Commissioner who had not been impressed by one of the major AI tools. The response it provided was only partially accurate. He took out his phone and used a version linked to verified source material to run the same query. The result was comprehensive, precise, and footnoted.

The technology itself isn’t particularly impressive. It’s the awkwardness that surrounds it. Businesses are purchasing licenses that their employees discreetly cease to use. At conferences, partners boast about adoption and then request that paralegals print materials. This exact pattern—good tools, poor workflows, and cancelled subscriptions—was detailed in a LinkedIn post that received hundreds of reactions a few weeks ago. It’s not that AI is ineffective. It’s because no one has figured out how to bill for twelve minutes of work that took an hour.
The CEO of LexisNexis, Sean Fitzpatrick, believes that top attorneys will see an increase in their hourly rates—possibly approaching $10,000—because AI will increase rather than decrease the value of their judgment. Perhaps. There is also a counter-story, the one that paralegals are currently experiencing, in which entry-level jobs that once supported a career are becoming less common. It’s difficult to ignore the possibility that both statements could be true simultaneously. The occupation won’t go extinct. However, the route to it is beginning to look very different than it did even three years ago, and no one seems quite prepared to publicly state this.









