I’ve been using AI in my product marketing work to review how software directories describe OnPage. I’ll give it a set of listings and ask it to flag places where the copy gets our product category wrong or leaves out something that matters to buyers, like on-call scheduling or alert escalation or alert noise reduction. That’s usually where I take over and decide what to change. Once i have the final copy, I simply ask it to log into the back-end of those directories and submit the changes. One day it struck me: what if I gave an AI assistant a set of rules and asked it to carry out the work inside our own product when I’m trying to create a dummy set of schedules inside OnPage ahead of a demo?
I picked a task our IT customers know well but would love to offload to someone else considering how much of a grunt work it is: building a 24/7 on-call rotation. I signed in to the OnPage Enterprise Web Management Console through ChatGPT’s cloud browser and described the coverage I wanted in plain English. I gave it six responders, separate schedules for daytime, weekday nights, and weekends, and a rule for when OnPage should escalate to the backup. Then I watched it build the schedules and check the Monday handoffs through the first six-week cycle.
In the video, I show how I used ChatGPT and OnPage to build a 24/7 IT on-call rotation with six responders, configure escalation to the scheduled backup after five minutes without a reply, and verify the rotation for gaps and overlaps.
Watch the on-call scheduling demo on YouTube.
Start with one team and one coverage policy. This could be an SRE team covering production outages, an MSP routing urgent tickets to its after-hours service desk, or an IT operations team responsible for a platform that runs around the clock.
Give your AI assistant the escalation group name, approved responders in rotation order, coverage windows, time zone, handoff time, and escalation rules. Be specific about what fair coverage means. In this demo, each person had one assignment per week and covered each of the six primary and backup roles once per cycle. If your team also needs to account for holidays, time off, or specialist coverage, include those constraints in the prompt.
Here’s a prompt you can adapt for your own team. Replace the placeholders with approved contacts and your start date.
In the signed-in OnPage Enterprise Web Management Console, create an escalation group called [GROUP NAME]. Add these six approved responders in rotation order: [A], [B], [C], [D], [E], and [F]. Set OnPage to escalate to the scheduled backup after five minutes without a reply.
Build a recurring 24/7 rotation starting Monday, [START DATE], at 9 a.m. Eastern Time. Create these three schedules and assign the primary first and the backup second:
- Day Shift — Monday through Friday, 9 a.m. to 5 p.m. Eastern. Week one: A is primary and D is backup.
- Weekday On-Call — Monday through Thursday, 5 p.m. to 9 a.m. Eastern the following morning. Week one: B is primary and E is backup.
- Weekend On-Call — Friday at 5 p.m. through Monday at 9 a.m. Eastern. Week one: C is primary and F is backup.
Every Monday at 9 a.m. Eastern, advance each assignment one position through the roster, wrapping from F back to A. Repeat the six-week cycle with no end date. Use the coverage type as the calendar label and list responder names in the schedule details.
After saving, verify all six weeks and the handoff back to week one. Check that each coverage window has exactly one primary and one backup, each person has one assignment per week, and consecutive coverage windows meet without gaps or overlaps.
The calendar check matters most at the boundaries. Day Shift hands off to Weekday On-Call at 5 p.m. Monday through Thursday. Friday’s Day Shift hands off to Weekend On-Call at 5 p.m. The weekend ends Monday at 9 a.m., when the next week’s Day Shift starts and all six assignments advance.
For help setting up these rules, see the OnPage recurring schedule guide and the OnPage escalation group guide.
A schedule tells you who should be available. An incident response workflow also needs to determine which events warrant a page, who gets it first, when the backup should step in, and how the team confirms someone took ownership. An AI assistant can help configure and review the workflow. OnPage applies the saved routing and escalation rules when an alert comes in.
| Incident response task | How I’d use AI and OnPage |
|---|---|
| Check coverage before a handoff | Ask the assistant to review upcoming coverage in OnPage and flag uncovered hours, overlapping assignments, or holiday exceptions that need review. |
| Define escalation behavior | Specify the primary, backup, escalation delay, and what stops escalation. OnPage escalation groups can use delivery, reading, or replying as the stopping condition. Choose the event that matches how your team confirms ownership. See the escalation group guide. |
| Test the alert path | Use the assistant to draft a test plan, then trigger an approved test alert. Verify that it reaches the scheduled primary, escalates to the backup under the intended conditions, and records the expected status. See the OnPage dispatching and status guide. |
| Connect paging to your incident tools | A qualifying ServiceNow incident can trigger an OnPage notification, and OnPage can relay notification progress and responder replies back to ServiceNow. Teams using Microsoft Teams can also acknowledge and respond through the OnPage integration. See the ServiceNow notification workflow and the Microsoft Teams integration. |
For this demo, ChatGPT worked directly in OnPage through its cloud browser. OnPage also offers an MCP Server for compatible AI assistants. With the appropriate tools and permissions, an assistant can identify the current on-call responder, analyze alert workload, initiate an approved page, and track acknowledgment or escalation.
Equal assignments can still produce uneven workloads. A Friday release or a recurring overnight job can generate more pages during one coverage window than another. I’d start with the OnPage dashboard for alert activity and average read time, then use OnPage reports to compare groups or shifts over a defined period.
Here’s a second prompt you can use with an approved OnPage report export:
Compare the last four complete weeks of OnPage alert activity across daytime, weekday night, and weekend coverage. Use Eastern Time to assign alerts to shifts. Where the export supports it, show alert volume, alerts per covered hour, read rate, time to read, delivery failures, and escalation rate. Include the counts and denominators, flag small samples, and suggest three things to investigate before changing the rotation.
Start with one escalation group, verify coverage through the first rollover, and run an approved test alert. Once you have enough activity to review, use the data to decide whether the next change belongs in the schedule, the escalation policy, or the rules that determine which events deserve a page.
When essential platforms become unavailable, the disruption extends beyond the IT department. Employees lose access…
What You Need Before You Configure Alerts A detected outage does not reduce downtime by…
It is late in the evening when the employee on call receives an urgent request.…
IT operations teams need more than a shared calendar to manage after-hours incidents. When a…
When a critical incident occurs, reaching the right person quickly can make all the difference.…
For a small practice, choosing an electronic health record system affects nearly every part of…