For most of the last thirty years, legal technology has been solving the same problem: information. How to store it. How to find it. How to retrieve it. How to move it around faster. The latest wave of AI appears to accelerate that trajectory dramatically — and in some areas, genuinely and unambiguously so. Workflow automation, eDiscovery, breach review, regulatory filing, document processing at scale: these are areas where GenAI and agentic AI are doing something the preceding twenty years of technology largely could not — dramatically reducing, and in some cases entirely replacing, human effort. Work within the law exists somewhere along a continuum defined by Art at one end and Science at the other. The examples above are the Science end. That is real, it is significant, and this series does not dispute it. What this series examines is the other end of the spectrum: the Art; the creative, judgment-intensive, contextually complex work that has always commanded the highest rates, resisted the most sophisticated previous tools, and on which the most ambitious productivity claims are now being made — the end AI has not yet meaningfully reached, and toward which it is advancing. Four articles. Four assumptions worth examining: how productivity in those areas is measured, how expertise is reviewed, how context is captured, and where judgment comes from.