24-Hour Triage Checklist

The daily routine of the notification triager. Fifteen minutes, first thing in the morning. Zero-inbox rule. Every notification routed, every operator heard back from within 24 hours.

Companion to: Fix the System, Chapter 5. Who uses it: the notification triager — usually a coordinator, sometimes the planner, occasionally the maintenance supervisor. Whoever owns this role owns it daily, not as an aside. Time required: 15 minutes per day, every working day. First thing in the morning is best. Do not triage at the end of the day; the 24-hour clock will lap you.

The triager's first fifteen minutes

1. Open the notification queue

• Open your CMMS or the shared notification form inbox.
• Filter to notifications created in the last 24 hours that are still in NEW / UNTRIAGED status.
• If the list is empty, you're done. Go.

2. For each notification, read and route

For each one, answer three questions in this order:

Q1 — Is this actionable? A notification is actionable if it describes a specific observation on a specific asset that could generate a work order, a PM review, or a planning input.

• If yes → go to Q2.
• If no (vague, duplicate, unrelated) → REJECT with written reason. Go to the next notification.

Q2 — Is this time-critical? A notification is time-critical if waiting more than 2-3 days significantly raises the risk of failure, safety, or production impact.

• If yes → IMMEDIATE ACTION route. Generate a WO now or today. Flag for dispatch.
• If no → go to Q3.

Q3 — Does this require a planned job? A notification requires a planned job if the work involves parts, procedures, coordination with operations, or safety lockouts that warrant formal planning before execution.

• If yes → PLANNING QUEUE route. Create a WO in planning's queue with the notification text as the starting point.
• If no (minor cosmetic issue, low-consequence observation) → REJECT with written reason, OR batch for the next minor-work window.

3. Close the loop with the operator

For every triaged notification (immediate, planning, or rejected):

• Update the notification status in the CMMS so the operator can see the disposition.
• If the operator included contact info, send a short one-line acknowledgment (a Teams/Slack message, email, or annotation on the notification form).
• For rejections, write ONE sentence stating why. Examples:
• "Rejected — duplicate of WO 47821 (same equipment, same observation, already in planning queue)."
• "Rejected — asset is on the Q3 replacement list; no interim action justified."
• "Rejected — observation is within normal operating range for this equipment class; no deviation."
• "Rejected — insufficient detail to act on. Please resubmit with more specifics about which valve and under what conditions."

Never leave a rejection blank. A silent rejection is a broken promise.

4. Log the triage metrics

At the bottom of the day's triage, note:

• Count of notifications triaged today: ___
• Count routed to IMMEDIATE: ___
• Count routed to PLANNING QUEUE: ___
• Count REJECTED: ___
• Any notifications NOT triaged within 24 hours? ___ (should be zero — flag if not)

What to do if the queue blows up

Occasionally the queue arrives at 40+ notifications in one morning. Usually after a shift handover where a new operator has noticed ten things at once, or after the first week of a new pilot when operators are catching up on things they've been silent about for years.

When this happens:

• Don't skip triage. Fifteen minutes becomes forty-five. Still triage everyone.
• Batch the rejections. If three notifications are duplicates, reject two with "duplicate of notification X" and keep one active.
• Batch the cosmetic/minor items. Create a single batch WO for minor-work cleanup during the next low-production window. Link all the minor notifications to that WO.
• Do not triage while tired. If you're triaging at end of day, stop. Come back in the morning. A bad triage creates more work than a one-day delay.

What to do if the queue is empty

If the queue is empty for more than two days running:

• Walk the floor. Ask operators if they have anything they haven't written up.
• Check the "I'll tell Mike" channel — talk to the walking coordinators, the shift leads, the mechanics. Are things being reported verbally that should be in the queue?
• Check the form. Is there friction — login issues, tablet not working, paper forms not being collected?

An empty queue is not a healthy queue. A healthy plant generates 2-6 notifications per operator per month. Do the math for your plant and see if the actual flow matches.

Weekly and monthly cadence

Weekly (30 minutes):

• Pull the week's triage metrics.
• Count by category (immediate / planning / reject).
• Flag any notifications that took longer than 24 hours to triage. Walk the floor and find out why.

Monthly (30 minutes, with operations + maintenance together):

• Walk through the month's reject pile.
• Align on what counts and what doesn't.
• Adjust the triage rubric if needed.
• This meeting is non-optional. It is the discipline that keeps identification installed. (Use the companion Monthly Reject-Pile Review Agenda for structure.)

From Fix the System by Ivan Getov. Companion to Chapter 5: Identify. getov.xyz/toolkit/fix-the-system/ch05/triage-checklist. Free. No email required.