Persistent Professional Agents: A Lifecycle Governance Thesis for Personally Controlled AI Across Organizational Boundaries

Persistent Professional Agents presents a lifecycle governance thesis for personally controlled AI agents that persist across organizational boundaries. It examines how an owner-associated agent can enter, operate within, and leave governed environments without persistent identity becoming persistent authority or legitimate access becoming unrestricted use. The thesis proposes five governing principles and develops a Host Participation Framework for admission, qualification, scoped authority, information rights, contribution, change, evidence, and termination. It introduces Context Settlement as a mechanism for reconciling authority, information, artifacts, contributions, custody, continuing obligations, and access when participation changes or ends, and a Contribution Compact for making reciprocal transfers of value explicit. The work is presented as a research-grounded product thesis and downstream design hypothesis rather than a legal doctrine, technical standard, or claim of patent novelty. Its central question is whether existing governance mechanisms can be composed into a practical, domain-independent lifecycle model for persistent AI agents operating across institutional boundaries.