On this page
The escalation ladder: when an executive assistant should interrupt the principal

An executive assistant protects the principal's attention, but protection is not the same as isolation. Some messages must reach the principal immediately, and some can wait until the next scheduled check-in. The difficult part is the middle: a client change, a board question, or a travel disruption that is urgent to the sender but not yet urgent to the executive. Without a shared rule, the assistant decides alone, and the executive experiences the result as either noise or delay.
An escalation ladder turns that judgment into an agreed sequence. It names the conditions that justify each level of contact, the response time expected at each rung, and the record the assistant keeps afterward. The ladder is not a sign of distrust. It preserves speed without asking one person to guess how much attention a situation deserves (Harvard Business Review).
Start by defining an interruption
An interruption is any contact that asks the principal to stop the current activity and respond. That definition includes a phone call, a direct message, a knock on a door, and a hallway conversation. It also includes a message that appears passive but triggers a reply within minutes. If the assistant sends a note marked urgent and then waits for the principal to notice it, the practical effect is still an interruption.
The definition matters because offices often count only the loudest events. A principal may report two interruptions in a day while an assistant has sent nine messages that each required a decision. The ladder should use the principal's experience as the unit of analysis, not the assistant's send button. Both people should measure the same thing.
Build four rungs, not a single alarm
Four rungs are usually enough. Each should have a name, a trigger, a contact channel, and a response window. The labels can be ordinary words: queue, alert, call, interrupt. Both people should recognize the rung without opening a policy document during a busy afternoon.
- Queue: the item can wait for the next scheduled review, usually the same day or the next morning.
- Alert: the item needs a decision within a defined window, often two to four hours. The assistant sends one written message with the decision, the deadline, and the default if no reply arrives.
- Call: the item is time-bound and a written exchange is slowing it down. The assistant calls once and leaves a short message if there is no answer.
- Interrupt: the item is urgent, material, and cannot wait for the current meeting or block to end. The assistant uses the agreed channel and states the reason in one sentence.
The rungs should not map to the sender's emotion. A counterparty who marks every message urgent does not thereby raise the rung. The assistant maps the item using the consequence, the deadline, and the principal's prior instructions. That distinction protects the principal from borrowed urgency.
Tie each rung to a response window
A rung without a response window becomes a suggestion. The window tells the assistant when to move up and tells the principal how quickly a message at that level expects attention. A queue item might expect a response by the next business morning. An alert might expect a decision within two hours during working hours. A call might expect a call back within thirty minutes. An interrupt expects the principal to stop, even if only to say "handle it."
The windows should reflect the principal's actual schedule, not an ideal one. If the principal is in back-to-back board meetings every Thursday afternoon, an alert sent at 1:00 p.m. may not receive a reply until 5:00 p.m. The assistant should either adjust the expected window or use a higher rung for genuinely time-bound items. Repeating an alert because the principal did not answer is a sign that the ladder and the calendar disagree. Record the expected window next to the rung.
Write the windows in plain language and keep them short. A one-page ladder that both people can recall is more useful than a detailed matrix that nobody reads. Review the windows after a month and change only the ones that caused a real miss. The next review will show whether the change held.
Write down the trigger, not just the category
Categories such as "client issue" or "legal matter" are too broad to guide action. A client issue might be a routine scheduling request or a threat to cancel a contract. The trigger should describe the observable condition that raises the rung. For example, "a top-five client asks to move a meeting that starts within four hours" is a trigger. "A client is upset" is a mood, not a trigger.
Good triggers combine three facts: who is affected, what deadline is at risk, and what happens if the principal does not respond. The assistant can then compare the item to the ladder without needing to interpret the principal's mood. This is especially important in remote support, where the assistant may not see the principal's calendar or the room they are in. The written trigger becomes the shared context that replaces proximity.
The ladder should also name the exceptions. A safety incident, a legal hold, a press inquiry with a same-hour deadline, or a family emergency may bypass the normal rungs. The assistant should know the bypass exists and should record why it was used. An exception that is never reviewed becomes the new normal, and the ladder quietly loses its authority.
Record every use of rung three and four
The highest rungs are the ones most likely to drift. An assistant who calls twice in a week may be over-escalating, or the principal may be missing messages that should have been caught earlier. The only way to tell is to keep a short record: time, trigger, rung used, channel, response, and outcome. The record should take less than a minute to complete, or it will not survive a busy day.
Review the record weekly, not annually. Look for patterns rather than individual mistakes. If most calls happen on the same day of the week, the calendar may be the problem. If most interrupts come from one counterparty, the relationship may need a different channel. If the assistant is consistently choosing a lower rung than the outcome suggests, the triggers may be too vague. If the principal is consistently ignoring alerts, the response windows may be unrealistic.
The record also protects the assistant. When a principal asks why a message arrived at an inconvenient time, the assistant can point to the trigger and the agreed rung. The conversation shifts from blame to calibration. That shift is the main reason to keep the record, even when nothing goes wrong. For a related method of transferring context, see how to hand off decision context.
The goal is not to eliminate interruptions. The goal is to make them deliberate. A working ladder gives the assistant a defensible answer to the hardest question in the role: should I interrupt now, or can this wait? For the handoff on the other side of that decision, see the acceptance check for remote executive handoffs.