DeepJudge's Agent Handoff Protocol Makes Context Portability a KM Infrastructure Question

DeepJudge launched the Agent Handoff Protocol (AHP) on August 13 — an open specification that allows users to move between AI platforms while carrying with them the objective, selected conversation context, documents, and analysis they have already completed. Harvey has announced it is building an AHP integration entering beta this month; Thomson Reuters has confirmed CoCounsel for Legal support. The protocol is built on top of MCP, not as a replacement for it: MCP handles discrete retrieval actions within a platform, while AHP moves the user and their task context into a different platform's complete interface. For KM directors, AHP is not merely a product feature — it is an infrastructure signal. If context portability becomes a buyer expectation, the quality and govern-ability of what travels in a handoff (firm precedents, matter profiles, playbook references) will determine whether AI output produced downstream is grounded in institutional knowledge or starts from scratch. KM must document its position on which knowledge assets are AHP-accessible and under what conditions.