Separate approval of a repeatable procedure from permission to use it today. Give each execution a clear eligibility check, owner and recovery path.
A team performs the same small service change every week. Sending each occurrence through a full approval meeting wastes attention. Calling all future occurrences “preapproved” creates a different problem: the environment changes, the procedure grows stale, and yesterday’s permission becomes today’s assumption.
A standard change workflow should connect two records: the approved change model and each execution of that model. The model defines the permitted activity; the execution proves that this particular use stays inside those boundaries.
This blueprint is a recommended operating design for IT-service change teams. It does not replace your organization’s change policy, authorize technical access or prescribe a production procedure. The example and proposed controls below are illustrative, not universal requirements.
Preauthorize a bounded procedure, not a category label
Atlassian’s change-type guidance describes standard changes as low-risk, repeated activities following an approved, documented process. The important distinction is that the procedure has already been assessed—not that somebody selected “low risk” on a new request.
Define a model narrowly enough that an implementer can answer whether a proposed instance matches it. “Routine administration” is too broad. A specific action on specified service components, using a known procedure under stated conditions, is testable.
Keep a normal review route for work outside that definition. If your policy requires a fresh discretionary approval for every occurrence, describe it as a streamlined approval workflow rather than assuming it is a preauthorized standard change. ServiceNow’s Modern Change adoption playbook distinguishes standard changes from low-risk models that retain runtime approval.
Give the model its own lifecycle
The reusable model needs an accountable owner, a procedure version, an authorized scope, an approval record and a way to stop future use. Do not represent all of that with an unchecked template attachment.
| Model state | Decision owner | Required transition evidence |
|---|---|---|
| Proposed | Model owner | Scope, exclusions, procedure, test and recovery approach, prior execution evidence and expected impact are ready for assessment. |
| Under assessment | Policy-defined change authority | Record approved boundaries and version, or return the proposal with a reason. Requesters cannot declare their own proposal active. |
| Active | Model owner | Only eligible instances may reference the authorized version. Record review due date and designated backup owner. |
| Suspended | Designated suspension authority | Record trigger, affected versions, time, reason and disposition of open instances. No new dispatch under the suspended permission. |
| Retired | Change authority | Withdraw from selection, resolve remaining instances and retain the historical model/approval relationship. |
Use your policy to name the actual change authority; it need not be the same committee everywhere. Boston University Engineering IT’s policy, for example, describes model-based preauthorization by its CAB. That is an institutional authority arrangement, not permission for another organization to copy its approval hierarchy.
NC State’s change-management process separates proposal approval from use of the resulting standard template and describes governance for modification or retirement. Use that distinction when designing your own records: editing a live model must not silently rewrite what an earlier instance was authorized to do.
Check eligibility again for each execution
Model approval and instance readiness answer different questions. The first asks whether a defined activity may follow the standard route. The second asks whether the present request actually qualifies and is ready to start.
Keep a small eligibility receipt
For each instance, retain the model ID and authorized version, target service/component, requested action and parameters, requester, implementer, planned interval, evidence checked and the resulting eligibility decision. Record who or what checked it and when. Missing information is “not established,” not a pass.
- Scope: target, action, scale and exclusions match the approved model.
- Authority: the model/version remains active and the assigned implementer is permitted to perform the work.
- Conditions: the current environment, dependencies and available recovery arrangements still match the model.
- Timing: the interval is permitted and does not conflict with a protected period or other work.
- Readiness: required context, notifications and verification responsibility are present.
These are proposed checks to configure under local policy, not a claim that every product enforces them automatically. Keep timing controls connected to the blackout-window workflow; preauthorization should not be interpreted as an override of a separate scheduling restriction.
Make dispatch an accepted handoff
The coordinator sends the implementer a complete instance with the pinned model version, eligibility receipt, execution interval, procedure reference, stop conditions and verification owner. The implementer accepts only after confirming the target and readiness. Until that acceptance is recorded, the coordinator owns the unaccepted work and follows up through the designated backup—not an unmonitored team inbox.
Recheck the time-sensitive conditions immediately before dispatch. A request that qualified when logged may no longer qualify after a model suspension, target change or new service incident.
Use an instance state model that cannot confuse “done” with “safe”
| Instance state | Accountable owner | Next transition |
|---|---|---|
| Logged | Coordinator | Complete the required inputs and identify one active model/version; otherwise return for clarification. |
| Eligibility pending | Coordinator or configured checker | Record a pass to reach Ready, or move to Held with a failed/unknown condition. |
| Ready | Coordinator | Confirm current eligibility and receive implementer acceptance before entering In progress. |
| In progress | Implementer | Follow the approved procedure; move to Verification after execution, or Recovery when a stop condition occurs. |
| Verification | Named verifier | Compare the observed outcome with the model’s acceptance criteria and retain the result. |
| Closed | Coordinator | Record disposition, evidence and any linked recovery work. Distinguish successful, unsuccessful and cancelled outcomes. |
| Held / Recovery | Named exception owner | Record cause, next decision and resume condition; never advance solely because a target time elapsed. |
Keep failed eligibility separate from failed execution. A request outside the model can be routed to the normal change-control process before anything happens. A change that has already caused an unexpected result needs an owned recovery decision, not a relabeling exercise.
Suspend the model without abandoning its instances
A newly discovered failure, changed dependency or invalid procedure can affect more than the record where it was found. Define which role may suspend the model, notify its owner and coordinator, and identify every open instance referencing the affected version.
- Not started: hold dispatch, retain the original authorization reference and assign a reassessment owner. Resume only when an authorized active model applies and eligibility is checked again, or route to normal review.
- In progress: notify the implementer and responsible service owner. Follow the procedure’s stop/recovery rules and incident process as applicable; do not assume an abrupt stop or automatic rollback is safe.
- Awaiting verification: preserve observed results and finish the necessary outcome checks. Model suspension does not prove that an already executed instance failed.
- Already closed: retain the historical result. Open linked follow-up work if the new finding requires investigation; do not rewrite completed records to imply they used a later version.
The model owner proposes the correction and supporting tests; the designated authority decides whether the revised model may become active. Record which queued instances need reassessment. Never treat reactivation as a blanket release of every old request.
Handle ordinary exceptions explicitly
A missing target identifier stays Held with the requester responsible for correction and the coordinator tracking a response deadline. Duplicate instances are linked and one is cancelled before dispatch. An unavailable implementer goes to the approved backup; if no permitted receiver accepts, work remains undispatched. Missing verification evidence stays with the verifier, with an overdue escalation to the service owner. None of these conditions is resolved by changing the label to “standard.”
Example: a description change that stops qualifying
In this constructed example, a team’s policy authorizes correction of the description field on one existing internal distribution list per instance. It excludes membership, ownership, permissions and routing changes. The model specifies the procedure, eligible targets, verification and recovery approach; this is not a technical instruction or a recommendation to classify all such edits as low risk.
The coordinator logs a spelling correction against model version B. Before dispatch, the implementer finds that the supplied target identifier points to a different list. The instance becomes Held, not In progress. The requester supplies the intended identifier, and the coordinator repeats the scope and readiness checks.
If the requester instead asks to add members, the request leaves the standard route for normal review. If an execution exposes a defect in version B’s procedure, the designated authority suspends B and the coordinator holds other undispatched instances. Correcting the one request does not resolve a model-wide defect.
Measure the route without rewarding misclassification
Review demand and flow together: incoming instances by model/version, Ready queue age, time waiting for implementer acceptance and time awaiting verification. Compare successful executions with started executions; report held, diverted and cancelled requests separately so they do not disappear from the demand picture.
Track failed eligibility by reason, unexpected outcomes, recovery cases, expired reviews and the time needed to account for open instances after suspension. A growing standard-change share is not success if teams are stretching the approved boundaries. Use repeated disqualifications to refine intake or propose a separately assessed model—not to weaken the gate.
Where Workhint fits
Workhint’s Workflow Engine describes routing requests and approvals to defined roles and requiring information or files before work advances. Those capabilities are relevant to proposal assessment, complete intake and accountable handoffs.
The Workflow Automation Software solution describes configurable forms, stages, owners, permissions, tasks and event-based notifications. That is a basis for discussing a model register and instance queues. It is not evidence of a built-in ITSM catalog, automatic production execution, immutable audit storage or automatic revocation of in-flight work. Confirm the required behavior before relying on it.
Build This in Workhint
Use the following as a proposed configuration brief, not a statement that every control is already available:
Design a standard-change coordination system with separate model and execution records. Let model owners propose procedures and named change authorities approve versions, suspend use or retire models. Requesters submit instance targets, parameters and timing; coordinators check eligibility; permitted implementers accept dispatch; verifiers record outcomes. Restrict model authorization and execution actions by role. Include the model and instance states above, required evidence, failed/unknown eligibility holds, duplicate detection review, backup ownership, overdue escalation and suspension review of open work. Provide active-model, ready-instance, held-instance and verification queues, with demand, queue-age, outcome and exception measures. Retain a history of version references, decisions, ownership changes and notifications. Confirm how version pinning, release rechecks, history retention and suspension controls can be implemented and tested; do not connect production actions until those requirements are validated.
FAQ
Does a standard change need a record for every execution?
In this design, yes: the instance connects the authorized procedure to the actual target, owner and outcome. Preauthorizing the model does not make execution evidence unnecessary. Follow your organization’s recordkeeping requirements.
Can a requester make an exception inside the standard route?
Not by changing the classification. When an instance falls outside the model, hold it and use the applicable review route. Only the designated authority can approve a changed model or another permitted authorization path.
What happens when the model review date is overdue?
Define the effect in policy before activation. If overdue review blocks use, dispatch must stop until the required review is complete. A reminder alone is not a release control.
The useful shortcut is a well-defined procedure with reusable authority. Keep that authority connected to current eligibility and truthful execution evidence, rather than letting a preapproved label become a permanent exception.

Leave a Reply