Legal AI has moved past the pilot stage. On October 16, 2026, Harvey announced that Hogan Lovells Cadwalader had adopted its platform across its practice groups and unveiled a firmwide AI strategy built on signed agreements with Harvey and Legora.
Miguel Zaldivar, CEO of Hogan Lovells Cadwalader, said, “Through Harvey, we’re perfecting the ways to capture and apply the expertise that exists across our firm on a variety of legal work.” The ambition is clear: turn a firm’s accumulated knowledge into systems that can support legal work across the business, rather than keep that expertise trapped in individual lawyers and old files.
The move reflects a wider shift in how firms are building AI. Instead of treating these tools as generic chat systems, law firms are creating systems around their own data, deal histories, trial strategies, and client preferences. That approach gives firms a way to apply AI to the work they already understand, while also making their internal knowledge part of the technology.
AI is becoming part of the legal workflow
Almost 90% of legal professionals in the U.K. and Ireland use AI. Almost 80% of firms using AI say they can handle more work without increasing resources, while over 70% say the technology cut costs by absorbing administrative work.
Other firms are putting the same tools into specific legal tasks. A&O Shearman has worked with Harvey to develop AI agents for reviewing loan agreements and analyzing regulatory filings. Slaughter and May has rolled out Harvey across all practice areas for regulatory research and document analysis.
These deployments place routine work at the front of AI automation. That matters because routine work has long occupied a large share of junior lawyers’ time, and those assignments also helped them learn how legal practice works. Remove the repetition and firms must rethink how lawyers build judgment, not just how quickly they produce documents.
The financial model is changing with the workflow. The share of work charged by the hour is expected to fall from 72% to 44% over the next two to three years, and about one in five firms that have widely adopted AI report difficulty meeting billable-hour targets.
Nick Rowles-Davies, CEO of Lexolent, put the problem in blunt terms: “You can’t charge 16 hours for something that takes 16 seconds.” The arithmetic was always going to notice.
Training, judgment, and the next legal career ladder
AI also changes what firms expect from lawyers. Firms are placing more value on curiosity and entrepreneurship, while building new career paths in legal engineering, product, and AI delivery.
Judgment is expected to become more valuable as lawyers spend more time checking and interpreting AI output. That shift comes with a hard limit: lawyers have been sanctioned for producing briefs with cases that don’t exist because of AI hallucinations.
The technology therefore does not remove the need for legal expertise. It changes where that expertise is applied, moving lawyers away from repetitive tasks and toward review, interpretation, and decisions about whether an AI-generated result can be trusted.
The workplace data adds another reason firms are pushing the change. 51% of legal professionals work evenings, but only 32% want to, while 22% work weekends compared to 11% who would prefer to.
“Lawyers [are] telling us that their day is getting better,” said Joshua Lenon, lawyer-in-residence at Clio. That does not erase the pressure around targets or the risks of poor output, but it shows why firms continue investing: AI can absorb administrative work, expand capacity, and remove some of the tasks lawyers never wanted to make their careers about.
The direction was already visible in the October 14, 2026, Business Insider article about firms building AI from their own data and the October 10, 2026, CNBC article about AI, billable hours, and legal careers. Hogan Lovells Cadwalader’s strategy makes the shift harder to dismiss as experimentation—AI is becoming infrastructure for legal work, with the billable hour and the traditional training ladder left to catch up.
Based on
