Korva is a scroll of the internet’s best ideas.
Legal tech stopped being a bad word
What changed for lawyers
The old objections were that lawyers could not be sold to, could not procure, and would not adopt new technology. Several things have changed at once. ChatGPT made AI mainstream, so lawyers use it in their own lives and their kids use it at school. New graduates arrive at law firms having done their coursework with AI, and they will not do the job without it.
Proof the market works
A few companies have also shown the market works. Harvey, Legora, and Crosby have raised money and found distribution, and they are proof that law firms are ready to buy this technology. Law has archaic, painful processes in every segment, from document drafting to discovery to divorce.
Lawyers become founders
The shift shows up in applications to startup programs too. Lawyers with 10, 20, or 30 years in practice are leaving high-paying jobs to build companies, because they have seen their peers use the tools and see that the work is manual enough to be changed. Bar associations have started running their first AI webinars on how to bring AI into a practice, as part of the continuing education lawyers must complete to keep their licenses.
Why was legal tech once hard to sell?
Many law firms have historically billed by the hour, so a tool that cuts hours can cut revenue unless pricing changes. Partners also carry personal liability for mistakes, which makes them cautious about unproven software. Buying decisions in firms are often made by committees of partners rather than a single procurement office. Those structural features slowed adoption of earlier generations of legal software.
Why do continuing education rules matter here?
Most US states require lawyers to complete a set number of continuing legal education hours to keep their licenses. Bar associations and other providers run the courses, so their topics signal what the profession treats as essential. Several state bars have also issued ethics guidance on lawyers using generative AI. When a topic enters required training, it moves from a curiosity to something everyone must understand.
Why does document review suit AI so well?
Discovery, the stage where each side hands over relevant documents, can involve millions of pages that people once read one by one. Software for sorting documents by relevance has been used in litigation since the early 2010s, and courts have accepted it. Language models extend this by reading text for meaning rather than matching keywords. The work is repetitive, text-heavy, and checkable, which fits what these systems do best.
More from Aparti
