IT systems · Practical guide

Validating ServiceNow for GxP: tracing an approved change to the resulting system state

A ServiceNow change is approved, its maintenance window met and the ticket closed. Do the actual settings match the approval, and does the GMP application still work as intended? Trace Change Management, the CMDB and implementation evidence together. Identify which questions the available standard records cannot answer.

Table of contents

At a glance

What do you need to validate in ServiceNow for GxP use?

Where ServiceNow controls GMP-relevant changes or evidence, assess the configured workflow. Link the approved scope, affected configuration, actual implementation, tests and closure decision. A ticket status, CMDB entry or supplier certificate alone is insufficient. Test alternative ways of changing data, enforcement of approval rules and audit coverage. Licence, release and local workflows delimit the available functions.

  • A change record links decisions to documentation; it does not replace the evidence of review at the affected system.
  • Proposed CMDB changes and their verification are not automatically a forced release gate.
  • An entry during the maintenance window does not establish that the change itself was approved.
  • Audit data, history view, and retained evidence have different functions.

1. Evaluate the specific ServiceNow process rather than the entire platform

This relates to ServiceNow Change Management within the ITSM context, its relationships with the Configuration Management Database, and the required platform documentation. This is not a general guide applicable to all ServiceNow products. Document your applications, license rights, activated extensions, integrations, and customized workflows. A supplier's security certificate may constitute relevant supplier evidence, but it does not automatically validate your specific release process.

The linked documentation shows the Brazil release reference. Individual enhancements are explicitly marked as Early Availability. This does not imply that these functions are generally available or activated in your instance. This article does not presume the presence of new lockdown, AI, or compliance functions as existing controls. Please verify your actual release and configuration status before converting a manufacturer's description into a requirement or test.

The regulatory reference pertains to the use of ServiceNow within the GMP manufacturing of human medicinal products. The key consideration is whether ServiceNow impacts relevant processes or records. A ticket for an office keyboard has a different impact than the approval of a LIMS interface change. Annex 11 provides a risk-based lifecycle framework for such scenarios; a universally "validated SaaS" cannot be inferred from its product name alone.

Sources: [7] [1] [8] [9]

2. Standard, Normal and Emergency professionally defined

ServiceNow distinguishes between standard, normal, and emergency changes. Standard changes are based on a pre-approved, repeatable process. Normal changes undergo evaluation and approval; emergency changes follow an expedited pathway. This describes the process mechanics, not the appropriateness for your GMP environment.

For a standard template, you should specifically define the authorized scope: affected system types, approved parameters, required prerequisites, and termination conditions. Any deviation outside these boundaries must not be permitted solely because a similar-looking template exists; instead, a suitable and intentionally inappropriate task must be reviewed. Additionally, clearly specify who is authorized to modify the template and what re-evaluation is required as a result.

For emergencies, authority, immediate decision-making, and subsequent completion of documentation must be clearly defined. Urgency does not constitute confirmation of technical soundness. A CAB may be part of your process; this article does not prescribe a universally required committee. What is essential is that the responsible decision-makers understand the relevant quality and data implications, and that the actual decision pathway remains traceable and justifiable.

Sources: [2] [9]

3. Compare CMDB relationships with the scope of changes

A Configuration Item describes a managed object. ServiceNow can assign multiple CIs to a single change and differentiate between impacted CIs and affected services. However, an existing link provides neither evidence of the completeness of your dependencies nor proof of the actual state of a server or application. For that, you require a suitable reconciliation process outside of simple ticket associations.

Start with the affected function: Which data pathway should be changed, and which application processes the result? Then compare the CI-list against architecture, responsible parties, and actual connectivity. The following workflow guide illustrates the required relationship. It does not prescribe separate documents or a specific ServiceNow data model; what is essential is a traceable connection.

Internal Work Aid: From Change Request to Reliable Conclusion
StepAppropriate contentWhat has not yet been demonstrated
Requirement Intended use, precise target values and affected functions That the change has actually been implemented.
Impact Affected CIs, Services, Data Paths, and Subject Matter Risks That every dependency was identified through an automated link.
Approval Authorized decision regarding the specific status and scope Later modifications remain fully covered by her.
Implementation Actual target with time, responsible party, and documentation evidence That the technical process operates correctly.
Verification Expectation, actual result, finding and assessment All other risks are automatically covered.
Closure Evaluated residual points and traceable usage decisions That a closed ticket status alone guarantees permanent suitability.

Sources: [3] [1] [9]

4. Proposed Changes do not constitute a prohibition on implementation

Under Proposed Changes, intended CI values may be provisionally applied and subsequently implemented. The ServiceNow documentation explicitly states that application is possible without verification or despite failed verification. Therefore, the need for a lock on a process cannot be inferred solely from the existence of this verification step.

Another boundary concerns the detection of planned changes. Documented standard logic can account for the temporal relationship with a change. ServiceNow highlights that this may result in a manual, contentually inappropriate change within the window appearing as planned. A suitable timestamp is therefore no substitute for comparison with the approved target value.

Define the required control specifically and assess its actual effectiveness. This could be a properly configured technical transition protection or another justified control within the overall process. Additionally, distinguish whether a change is occurring in a CMDB record or in the actual target application. An Apply action on the CMDB is not automatically equivalent to a deployment on the affected system.

Sources: [4] [9]

5. Example: an approved interface change

The following example is not a completed ServiceNow test. Change C-17 pertains to Gateway G1 and the two applications A1 and A2. Approval is granted for the change of an interface port from 8443 to 9443 in all three affected configurations. After the change, a predefined message with Batch L17 and result 7.25 mg must arrive intact and unchanged at both receiving endpoints.

The intended use requires prior legitimate authorization, an appropriate CI-list, and a successful technical pre-assessment prior to implementation. For the CMDB application, a technical lock is specifically mandated upon failed verification; this is not a claimed standard feature of ServiceNow. Negative pathways are separately constructed from the positive workflow and subsequently reset upon review.

The window 20:00 to 21:30 covers 90 minutes. 15 minutes of backup, 20 minutes of implementation and 25 minutes of testing are planned, a total of 60 minutes. In addition, 20 minutes of relapse reserve, which leaves ten minutes of buffers. The invoice shows only the planning consistency. Whether these times and the relapse procedure are suitable, an actual project would have to prove separately.

Expected results for separate checks of approval, content and time
Test caseStarting point in the exampleExpected result and supporting evidence
Scope is missing Only G1 and A1 are assigned; according to the architecture, A2 belongs to it. No final assessment is made until A2 has been taken into account and its impact has been evaluated.
Verification negative Pre-qualification fails; attempt is made to apply the CMDB change. Selected technical lock prevents application; otherwise, documented finding.
Incorrect content in the window At 20:25, instead, 9553 will be entered at G1. Detect a deviation from 9443, even if a time window check is flagged as "planned".
Positive flow 15 + 20 + 25 minutes for safeguarding, implementation, and verification. 21:00; 30 minutes remaining, comprising 20 minutes contingency and 10 minutes buffer.
Technical Result Comparison Message L17 / 7.25 mg via the new connection to A1 and A2. Both recipients demonstrate correct identity, value, and unit; no silent change.
Professional qualification fails Ticket indicates implementation completed, but A2 does not receive any notification. No successful overall closure; assess the finding and determine the anticipated rollback decision.

Sources: [4] [1]

6. Combine actual implementation and test evidence

Define which documentation demonstrates the actual target state. This could, for example, be a controlled configuration export, an appropriate deployment protocol, or a verified system status report. The selection depends on the target system. A detailed free-text entry in a change record does not serve as adequate evidence if it merely describes the intended action.

Attach the record to the approved version. If a parameter was altered, an application was omitted, or an additional component was modified during the process, the deviation requires evaluation. This does not automatically necessitate a complete new project; however, it must remain clearly evident which portions of the original evidence remain valid and where a new determination or review is required.

The closure should summarize the actually relevant results: What was implemented, which tests were successful, which findings remain unresolved, and who determines usage? A system error is not considered resolved merely by closing a task. Similarly, a single green test only validates its own specific objective. Define within the procedure how a sound overall conclusion can be derived from individual evidence.

Sources: [1] [10] [9]

7. Check roles, approvals and alternative ways to update records

Review who initiates a change, performs the technical evaluation, grants approval, implements, and finalizes it. Additionally, consider administrators, interfaces, and automated workflows. The documented field editability depends on the current status and configured workflow. Therefore, a read-only form view alone does not constitute sufficient evidence that all other relevant edit paths are adequately controlled.

As a targeted negative test, a change to the approved target value by a user with insufficient authorization is appropriate. Additionally, verify whether an automated import or technical service could alter the same situation. The test scope should follow the actually used and risk-relevant pathways. It is not justified to test integrations that your instance does not possess.

A license requires a clear subject and an authorized individual. Whether additional requirements for electronic signatures apply depends on the intended process and regulatory domain, not on the button label. Define how identity, decisional significance, and associated status remain aligned. A shared group identity must not inadvertently substitute for the required assignment to the decision-making person.

Sources: [1] [9]

8. Audit data, History Sets, and retained evidence are distinct

ServiceNow records auditing for activated tables and fields. Settings are not automatically propagated from a parent to a child table. The documentation also identifies exceptions, including certain upgrade and import paths. Therefore, verify the actual relevant fields and change channels in your instance, rather than concluding complete coverage based solely on an activated audit function.

History sets serve as a representation of audit data generated as needed. ServiceNow advises against using them as a basis for reports. Therefore, a clear traceability path displayed in the interface does not necessarily confirm that a long-term required audit trail can be fully exported, retrieved, and read. Select the intended reporting or retention approach based on the required documentation function.

For a relevant change in the target value, an auditor should be able to understand the previous and current situations, the responsible party, and the connection to the decision. Define additional appendices and external documentation requirements. Storage shall be determined based on purpose and applicable requirements. This article does not assume unlimited retention of all platform events nor a uniform deadline for all tickets.

Sources: [5] [6] [10]

Further reading:

9. The Change System continues to manage further development

If a workflow gets new permissions, writes an interface differently or updates a plugin, the reliability of the detection path can change. Therefore, not only evaluate updates of the controlled GMP applications, but also relevant changes to the ServiceNow process itself. Supplier documents and release notes help narrow down the scope concerned; they do not replace matching with your configuration.

For regression testing, specific critical transition points should be considered: lack of authorization, change in scope, negative verification, unauthorized field modification, and absence of a closure document. Even previously unchanged steps may be affected by a new rule. Existing validation evidence may continue to be utilized, provided its suitability for the current state remains clearly justified.

In addition, assess system operation during failure or limited access. How is an urgent change controlled and decided upon when the normal workflow is not available? How are related records restored to the intended audit trail following recovery? Such a procedure should align with your risk assessment. It is neither a blanket authorization for implementation nor a mandatory requirement for a second fully dedicated platform.

Sources: [7] [9] [10]

Further reading:

10. Preparation and Checklist for QA and IT

For a smaller company, a narrowly defined pilot process is often more helpful than a flat-rate platform list. Choose a recurring change with real GMP effect and a known data path. The Process Owner explains decisions and exceptions; IT provides instance, CI and integration knowledge; QA evaluates quality impact and evidence requirements. The implementation partner should be able to explain the actual effective rules.

Effort increases when CMDB ownership is unclear, workflows differ substantially, changes are hard to trace, evidence is stored externally or representative test data are missing. Start with a current workflow and identify its evidence gaps. The checklist supports that preparation. It is not a universal certification checklist.

  • Applications, license scope, release, and relevant modifications recorded.
  • GMP-relevant change types and justified exemptions defined.
  • Approved content linked to affected CIs and data pathways.
  • Verification and actual implementation of required controls differ.
  • CMDB status with real configuration and implementation documentation compared.
  • The review covers roles, technical accounts and alternative ways to update records.
  • Expert evaluation of technical assessment results and pending findings in the closing assessment is reviewed.
  • Audit coverage and long-term readability of required evidence verified.
  • Procedures for updates, emergency, and failure responses have been assigned to responsible personnel.

Sources: [1] [3] [9]

Project experience and collaboration

Compare the approved change with its actual implementation

With change owners and IT, we walk through a prepared configuration change. We compare approved target values and affected components against the configuration actually active afterwards. A standard path may need a few targeted cases. Alternative write paths and custom workflows need additional controls. The team verifies that the test evidence and closure decision belong to the specific change. We document who can make each decision and how a deviation is handled.

Frequently asked questions about ServiceNow

Must ServiceNow be GMP-validated for each IT ticket?

No. The key consideration is whether the specific use impacts GMP-related processes, decisions, or records. A general office support function differs from a binding change workflow for a LIMS. Justify the distinction and evaluate shared functionalities, such as permissions or integrations, for relevant implications. This enables the scope to be aligned with actual usage, without either including every platform function or indiscriminately excluding ITSM.

Sources: [9]

Does a ServiceNow security certificate fulfill customer validation?

No. A certificate can attest to a specific supplier performance or a defined control area. You must verify whether its scope aligns with your supplier assessment requirements. It does not automatically demonstrate that your customized change workflow involves the correct personnel, accounts for negative reviews, or verifies actual target achievement. Therefore, link appropriate supplier evidence to the specific requirements and validation data of your actual application.

Sources: [9]

Can a failed proposed changes review prevent system application?

A customer-specific control may address this. However, the documented standard procedure allows application even in the absence of or despite failed verification. Therefore, carefully evaluate the required locking effect and its relevant write access controls. If your technical requirement permits an alternative control, its suitability must be justified. A technical protective effect must not be inferred solely from visible inspection results.

Sources: [4]

Why is labeling as a planned change insufficient?

Since temporal assignment and substantive approval are distinct concepts, ServiceNow indicates that a manual CI change within the change window may be recognized as planned under standard logic, even if its content does not align. Therefore, compare the actual value against the approved target state. In the example of this article, port 9553 would be incorrect, even if it was entered during the intended window rather than outside of it.

Sources: [4]

Is a CMDB update sufficient as evidence of a successful implementation?

No. It initially describes the factual situation stored in the CMDB. For the actual current status, you require appropriate evidence from the affected system or from a reliable implementation pathway. Thereafter follows a technical assessment. In the example presented in this article, both applications must receive the correct message; merely changing a CI property to 9443 is insufficient to prove this. Therefore, link the CMDB entry, implementation, and outcome instead of replacing any one of these steps.

Sources: [3] [1] [9]

Can standard changes be used without requiring a new individual authorization?

The documented standard process is subject to prior approval. For its application, it must be clearly defined what repeatable scope it encompasses and what prerequisites must be met. A modified application, a different target parameter, or an open finding may fall outside the boundaries of this scope. Such cases must be deliberately assessed. The term "standard change" does not imply that any repetition is without critical evaluation nor does it override your established quality decisions.

Sources: [2]

Does every change necessarily require an electronic signature?

No, Annex 11, Section 14, does not establish a requirement to electronically sign every change. The necessary authorization derives from the process, the purpose of the record, and the applicable regulatory framework. When electronic signatures are used, it must be ensured that their meaning and enduring linkage to the relevant record are secured. Furthermore, a relevant authorization must have a clear subject matter and be issued by an authorized individual. Verify this alignment against the actual change process.

Sources: [9]

Why is the history view insufficient to address every audit question?

The view is a representation whose technical foundation and completeness require verification. ServiceNow describes History Sets as being generated as needed from audit data and advises against their use for reports. Furthermore, change tracking depends on table, field, and write path modifications. Therefore, verify the required evidence to ensure that the appropriate changes are included and can be retrieved later in the required context.

Sources: [6] [5]

What must be retested after a ServiceNow update?

Evaluate the changes to your app, configuration, and integration that are actually used. Especially relevant are transitions that affect permissions, CI assignment, change enforcement, or evidence. Regression can also include unchanged, indirectly affected functions. A universal full-retest follows as little as a waiver of customer tests. New functions in an early availability documentation are also not proof that they are already available to your productive instance.

Sources: [7] [9]

What documentation is required for a small QA/IT team at the outset?

Useful starting material includes a bounded real change, affected configuration items and data paths, approval rules, and representative implementation and test records. Include the instance version, adaptations and owners of evidence repositories. A specific question, such as an untraceable parameter change, helps identify whether the procedure, configuration or evidence trail needs work first.

Sources: [1] [3] [10]

Sources and review date

Source date: 6 October 2026. The linked ServiceNow documentation shows the Brazil release reference; individual enhancements are marked as Early Availability. Availability, licensing scope, and instance configuration are to be verified separately. Regulatory reference: GMP for human medicinal products. Figures and test cases are standalone work aids.

  1. Create a change request
  2. Change types
  3. Create a change from a CI
  4. Managing proposed changes
  5. Exploring Auditing
  6. Knowing about History sets
  7. Change Management release notes
  8. EudraLex Volume 4 – Human Medicinal Products – GMP
  9. EU GMP Annex 11: Computerised Systems
  10. EU GMP Chapter 4: Documentation

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Clarify the ServiceNow approval pathway for GMP-related changes

Daniel Herrmann Consulting supports you in aligning requirements, effective controls, and existing documentation of your change process.

  • Cloud and SaaS validation →

    Maintain the validated state of cloud and SaaS systems as suppliers issue new versions. We define the boundary between your responsibilities and the supplier’s and agree how each release will be assessed.

Send a project enquiry Customer testimonials →