ServiceNow Go-Live Checklist: What to Verify Before Launch

Go-live is the riskiest moment of any ServiceNow project. This readiness checklist covers the things teams most often forget — the ones that cause a rough launch. Run through it in your go-live readiness review and don't launch with open critical items.

Environment & configuration

• All required update sets / scoped apps are committed to production, in the correct order
• No update set conflicts left unresolved
• Production system properties reviewed (email sending, glide.servlet.uri, instance name)
• Cloned from prod recently enough that test mirrors reality (if applicable)
• Plugins activated in production, not just in dev/test

Security & access

• ACLs tested by impersonating real user roles — not just as admin
• Roles and groups populated correctly from the authoritative source
• SSO / SAML works in production and has a tested fallback login
• Admin and security_admin access is limited to the right people
• Guest/self-registration behavior is intentional, not accidental

Email & notifications

• Outbound email is enabled in prod (it's off by default on new instances)
• Inbound email actions tested with the real mailbox
• Notifications point at the right recipients (no test addresses left in)
• Notification volume is sane — no accidental storm on cutover

Data

• Final data migration completed and record counts reconciled
• Spot-checked migrated records for quality and correct references
• CMDB populated and Discovery scheduled (if in scope)
• Knowledge articles published (not left in draft)

Integrations

• Every integration tested in production with real endpoints and credentials
• MID Server (if used) is up, validated, and monitored
• Error handling and retry behavior verified
• API rate limits and credentials confirmed for prod

Performance

• Key lists and homepages load acceptably under realistic data volume
• Heavy business rules / scripts reviewed for efficiency
• Scheduled jobs won't collide at peak times

People & process

• End users trained; quick-reference guides distributed
• Fulfillers / agents trained on their queues and workflows
• Support model defined — who fields issues in week one, and how
• Communications sent: what's changing, when, and where to get help

Cutover & rollback

• Detailed cutover runbook with owners and timings
• Rollback plan documented and agreed (what triggers it, how to execute)
• Go / no-go decision criteria defined and a decision meeting scheduled
• Hypercare team and hours confirmed for the first 2–4 weeks

The items teams forget most

In our experience the top go-live surprises are: outbound email still disabled in prod, ACLs only tested as admin (so real users can't see their records), integrations only tested in sub-prod, and no rollback plan. Verify those four specifically.

Want a second pair of eyes before you launch?

A pre-go-live review by someone who's launched ServiceNow before catches the expensive misses while there's still time to fix them. If your team is approaching go-live without deep ServiceNow experience, a short readiness review with a specialist is cheap insurance against a painful launch.

Run pre-go-live checks with the ServiceNow Agent

Before you flip the switch, the NovusAI ServiceNow Agent can sanity-check your instance for you: ask it to look for inactive-but-referenced records, missing configuration, data-quality gaps, or specific patterns across tables, and it queries live and reports what it finds. It's read-only by default, so it's safe to point at a pre-prod instance for a final sweep. See how it works.

How the NovusAI agent does this

Most of this checklist is an audit, and an audit is a series of queries someone has to remember to run. The agent runs them:

• Security — which ACLs exist, which tables are unprotected, which accounts hold admin, whether any ACL was left open during testing.
• Update sets — what is captured, what is still open, what will not travel because the platform treats it as data rather than customisation.
• Performance — business rules without conditions, queries inside loops, rules firing on every update, the patterns that only hurt at production volume.
• Data — orphaned records, missing references, fields nobody populates.

Ask it: "Audit this instance for go-live risks" and it will report findings against real records, with the sys_ids, rather than restating the checklist you just read.

It will not sign off your go-live. But it removes the excuse that nobody had time to check, and it finds the items teams most often skip because they are tedious rather than difficult.