Legal AI deployments concentrate on document-intensive workflows where speed and recall are both economically and strategically important. Law firms and corporate legal departments face the same economics: large volumes of documents, high billable-hour costs, and high error stakes. The highest-impact applications are in e-discovery (review volume reduction), contract analysis (risk identification at scale), and legal research (precedent surfacing). Large language models have substantially changed what is possible in legal NLP since 2023, but hallucination risk requires human review workflows rather than autonomous output.