Ask. An assistant that answers from your program's data.
Ask a question in plain English about your schedule, your rules, or how to use the app.
This is not a general chatbot. Before it answers, it reads your program's own data: the rules, the
current schedule, past runs, the audit trail, and the app's own code and documentation. It can check a
proposed rule against the schedule, and it can run a test solve to see what would change. It saves
nothing.
Asked whether MICU could hold 3 residents a block instead of 2, it ran a one-minute test solve with the
new count in place of the current rule, and reported that every other hard rule still held and which
rotations the 13 extra MICU blocks would come from.
Every answer cites what it read: rule ids, run ids, cells. A second reader checks each draft before it
posts. A draft the second reader does not accept is held for a person to review.
It only reads. It cannot change a rule, a cell, a person or a setting. A grounded answer takes a few
minutes, not seconds.