Frank Childer – Trip report from the AI for Legal Reasoning and Adjudication Conference, Northwestern Pritzker School of Law, April 2, 2026 “…The Skill Gap Is Not About Prompting — It’s About Context and Oversight – The skills the legal profession needs for agentic AI are different from prompt engineering skills. They cluster around two capabilities that were largely absent from the day’s conversation: Context engineering is the practice of designing the information environment in which an AI system operates. What does the agent know? What sources can it access? What constraints shape its decisions? For a lawyer deploying an agent to review contracts or conduct discovery, the critical decisions are not about how to phrase a query: they are about what the agent is allowed to do, what it is required to verify, and when it must stop and ask a human. Critical evaluation of AI reasoning is the ability to assess not just AI outputs (is this citation correct?) but AI decision paths (why did the agent choose this approach? what did it not consider? where might it have been anchored by its initial framing?). This is closer to supervising a junior colleague than to spell-checking a document. It requires understanding enough about how the system works to recognize when its confidence is not warranted…”