Several university agents need the student handbook. The current file sits in a private team most agents cannot read, and a stale public PDF is what their knowledge sources index. What should the architect do with the data?
Select an answer to reveal the explanation.
Short Explanation
Five clerks with five different binders will quote five different handbooks. You publish one current copy to a place every approved agent can read, and you stop indexing the stale public PDF. Pasting the text into each agent's instructions just creates five more binders.
Full Explanation
Business data that more than one AI system needs should be published once, kept current, and shared under access that matches each consumer. The handbook is that data. A governed location with the current file, and permission for each approved agent, lets every agent ground on the same revision and lets the private team stay private. Pasting the handbook into agent instructions copies it into places that do not update when the registrar publishes a change, and it mixes policy text with the instructions that are supposed to control the agent. Fine-tuning a model per department freezes whatever PDF that department held and makes the next handbook revision a training project. Opening the private team to the entire student body widens access to everything else in that team so that agents can read one file, which is the wrong trade. Exam caveat: availability for other AI systems does not mean anonymous access. The governed location can still be closed to people who should not read the draft. Operational check: update one handbook paragraph in the governed location and confirm each approved agent returns the new paragraph and none of them still cite the public PDF.