List the Intraday metrics available to query, with definitions and units.

Resolve a user-friendly name (e.g. 'billing', 'GE') to a canonical queue.

Return current Intraday state for all (or selected) queues and the highest-risk queue.

Return per-interval forecast vs actual volume, AHT, and staffing for a queue.

Explain why a queue is at risk: ranked risk drivers with supporting data.

Identify queues with spare capacity that could help recover the target queue.

Simulate moving N agents from a source queue to a target queue for a duration (minutes).

Return a plain-English recovery recommendation (what / why / action / impact / trade-offs).

Replace the Intraday dataset with the user's own queues, so every other tool analyzes THEIR data instead of the built-in demo.

Each queue object requires: id, name, sla_current, sla_target, volume_forecast, volume_actual, aht_forecast, aht_actual, staffing_planned, staffing_actual (>= 1), occupancy, backlog, risk_status ("at_risk" | "watch" | "healthy"). Optional: aliases (list of strings) and intervals (list of objects with interval, volume_forecast, volume_actual, aht_forecast, aht_actual, staffing_planned, staffing_actual). Percentages are 0-100; times are seconds; staffing is agent counts.

Note: this sets the dataset for the whole server process, not per-conversation. Call reset_intraday_data to restore the demo dataset.

Restore the built-in demo Intraday dataset (undo load_intraday_data).