how to auto-close stale servicenow tickets safely

TL;DR

Define stale as no requester or assignee update for N days in a waiting state, send a 48-hour warning comment, then close as resolved for no response. Exclude incidents, VIPs, and anything with linked problems or changes. Log the rule name in the work notes so every closure is explainable and reopenable.

Steps

1. Write down the stale rule the team agrees on: no update from requester or assignee for 14 days while in Awaiting User Info or On Hold. Expected: a one-paragraph policy, not a tribal understanding.
2. Build it in Flow Designer or as a business rule: when a ticket matches the rule, post a warning comment giving 48 hours, then close as Resolved with close notes naming the auto-close rule. Expected: a test run on 5 real tickets posts the warning before any close.
3. Add exclusions: P1 and P2 incidents, VIP users, tickets with linked problems or changes, HR and legal categories. Expected: a dry run closes zero excluded tickets.
4. Log every auto-close with the rule name in the work notes and allow reopen by reply. Expected: a user who complains gets a clear explanation and a one-click reopen path.
5. Monitor for two weeks: count reopened auto-closes. Expected: reopen rate under 10 percent. Higher means the threshold is too aggressive; lengthen it.

Use this when

• The queue grows with tickets waiting on users who never reply
• Open-ticket counts never drop and leadership asks why
• Tier 1 spends mornings nudging dead tickets

Not for this skill when

• Active incidents or outages
• Tickets awaiting the IT team (that is our SLA problem, not the user's)
• Regulated tickets that require human closure

Compatibility

• ServiceNow: Flow Designer or classic business rules
• The same pattern ports to Jira Service Management automation rules

Variants

Users complain about closures

Lengthen the warning to 5 business days and offer reopen by email reply. Most complaints are about surprise, not the close itself.

Resolved tickets that never close

Use a separate resolve-to-close timer (for example 5 days after Resolved) rather than stretching the stale rule.

Why it happens

Tickets wait on user input that never comes, and nobody wants to be the person who closes them. Automation removes the social cost of closing a dead ticket.

Edge cases

• Tickets awaiting a vendor: exclude them and track vendor waits separately.
• Recurring or seasonal tickets that look stale but are not: exclude by category.
• Auto-close loops: exclude system comments from the activity check or the reminder itself resets the timer.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_vaqxVN9gAHwHiCYDqjxR3A