IT systems · Practical guide

Power Apps validation: low-code applications in GxP

Disabling a button may leave the underlying data accessible. A user who can write directly to a list or table must still be subject to the intended permissions. Identify the layer that actually enforces each critical control.

Table of contents

At a glance

Does a Power Apps application used in GxP need validation?

Yes, where the Power Apps application supports GMP-regulated activities. Annex 11 calls for application validation, and low-code applications are not exempt. Test depth depends on risk. Assess the app, data sources, formulas, connectors and permissions together. An application outside GMP scope does not require this GMP evidence, and other GxP domains have their own requirements. Validate the configured application for its use, rather than the whole Microsoft platform indiscriminately.

Scope and Applicability

Canvas- and model-driven applications, as well as various data sources, possess different control options. The following review questions serve as a starting point; specific product functionalities are not assumed for each variant.

Sources: [1] [2]

The app interface is only part of the system.

Formulas and compound specifications can influence technical decisions. Verify mandatory fields, limit values, status changes, and the handling of incomplete return entries. An empty results list may indicate that no data exist, but it may also indicate a failed or restricted access. These conditions must not be treated equivalently without proper consideration within the process.

Sources: [1] [2]

Describe the intended use of the specific app

Assess the application, not the label "Low-Code." An app for room reservation has a different impact compared to an application that captures test results or alters material status. Describe the data, decisions, and potential consequences of incorrect processing. This leads to an appropriate testing and control scope.

Capture app type, environment, data sources, connectors, embedded workflows, and actually used identities. A user action may trigger a background execution through another connection. Therefore, identify where authorizations and business rules are actually enforced.

Microsoft describes ALM for Power Platform through requirements, development, testing, change, deployment, and operations. Solutions and additional tools can support these activities. For your GxP process, document the tested app version including environment variables and relevant connections. It must be clearly traceable within the selected model which components and settings are actually transferred and which data source is accessed in production.

Sources: [1] [2]

Derate requirements from data and decisions

Define requirements such that their impact can be verified. "The app is secure" remains too general. "A production user can capture a test result but cannot modify the approved limit value" describes a specific boundary. Supplement the handling of missing inputs, repeated actions, and an unattainable data source.

Distinguish screen-level input validation from rules implemented in data processing. If a data record can be modified through another application, an import, or a workflow, screen-level validation does not cover all possible pathways. The architecture should clearly indicate which control applies to which pathway. Unclear alternative routes remain visible in the risk assessment.

Internal Work Aid: Derive requirements from data and decisions
ScopeQuestion to askOwn testing method
Identity Under which identity is reading and writing performed? Check roles and links with actual actions
Input What values and states are acceptable? Normal, boundary, and error cases are used
Processing Which rules modify the dataset? Compare results independently against the specialist guidelines.
Exception What occurs during a timeout or absent source? Controlled Interruption and Resumption
Traceability Which application instance generated the record? Version, data reference, and required history reconstruct

Sources: [1] [2]

Example and inspection table: Save via app, flow, and data source

An app records a test result for sample P-55. Due to delayed feedback, the user clicks "Save" twice. The own requirement mandates a clear, traceable record or a clearly identified repeat attempt. The document assumes no automatic default solution is implemented in any Power App.

Prepare the sample and the expected value. Perform a normal storage procedure and compare the data source with the displayed value. Then repeat the action under controlled delay. Verify the actual data inventory, not just the confirmation message. Two identical displays may represent two different data sets in the background.

Add a break following the writing but prior to the response, provided that such a scenario can be replicated in an appropriate test environment. The team must understand how an ambiguous outcome is identified and resolved. Blind re-saving may result in duplication. An appropriate control could involve a verified business key and a regulated retry mechanism.

Additionally, maintain that the flow is processed under which specific connection, and which app, flow, and configuration state is actually active in the target. Upon changes, the audit trail remains bound to this overall state. The same app name does not confirm an unchanged architecture.

Internal Work Aid: Save, Permissions, and Retry
Test caseActionExpected process outcome
Normal entry P-55 with valid value save Correct value assigned exactly to the intended operation
Missing mandatory specification Save without necessary identification No unclear release-related record
Repetition Controlled save with multiple triggers Selected repetition rule applies
Connection error Temporarily interrupt data source Error is visible, data status will be clarified
Unauthorized change Change critical parameters with subject-matter expert input Access corresponds to the specified limit
Technical Identity Capture actual flow composition The permission is linked to its effect on the process
Provision Compare app, flow, and relevant target configuration The verified status is actually active

Sources: [1] [2]

Control environments, compounds, and distribution

Separate development from intended productive use appropriately. A test application may leverage different data sources or permissions. Document these differences and assess their implications. Passing a test using a developer account does not validate the correct functionality with subsequent domain users.

Upon availability, it must be confirmed that the correct App version, associated flows, and required configuration are present in the target environment. Connections and environment values require particular attention if they reference other sources or identities. Successful import of a solution does not confirm the technical end-to-end workflow.

Obtain a traceable state justification. This includes the actually required components and their interrelationships. A change to a common flow can impact multiple applications. Therefore, impact assessment must extend beyond individual interfaces. Not every platform change requires the same test scope; the affected usage determines the selection.

Sources: [1] [2]

Operate, own, and prepare for fault investigation

Name the technical and professional responsibilities, as well as the designated representatives. Assess what occurs in the event of a personnel change or when the original developer ceases involvement. Personal dependencies and ambiguous ownership may complicate system operation. Therefore, the handover process should include consideration of access rights, system configuration, and required knowledge.

Define the handling of erroneous records. Who identifies an incomplete data capture event, who assesses the impact on the process, and how is correction implemented? A technical repair must not inadvertently obscure the original state. Traceability and retention must align with the specific data process.

For the next release, a small, well-kept set of critical regression tests is recommended. It should cover important entries, rights, processing rules and error paths. A pure click test of the surface is weaker than a comparison of the actual dataset and its usage. The selection is re-evaluated in case of changes.

Sources: [1] [2]

Next steps

  • Describe the intended use and critical decisions.
  • Define app, data retention, connections, and roles.
  • Check normal, boundary, and error cases throughout the entire chain.
  • Versioned release and controlled changes management.
Further reading:

Project experience and collaboration

Trace a saved record across all participating components

With the process and app owners, we follow a transaction from data entry through the flow and connection to storage. We interrupt the connection under controlled conditions to see whether saving again creates a second business transaction. The architecture determines whether an existing business identifier is sufficient or another restart rule is needed. We check permitted actions, error handling and the deployed configuration as a whole. The user interface is where the investigation begins.

Frequently asked questions about Microsoft Power Apps

Does Dataverse make the app automatically GxP-compliant?

No. Platform functions can support controls. Their configuration, usage, and suitability for the specific process must be evaluated.

Sources: [1] [2]

Can the department develop the app itself?

Yes, if competencies, responsibilities, and controlled development and release processes are in place. Simple usability does not replace these requirements.

Sources: [1] [2]

Must every optical change be revalidated?

The implications should be assessed. A simple color change and altered mandatory field logic do not necessarily require the same level of testing effort.

Sources: [1] [2]

Is Low-Code automatically less validation-intensive?

Not automatically. The scope is determined by intended use, risks, and complexity. A simple interface can integrate multiple data sources and workflows. Appropriate platform justification can support, but not replace, the evaluation of the own configuration and processes. Crucially, it is essential to identify which specific business decisions and records are influenced by the particular application.

Can we simply re-publish an app after modifications?

For a GxP-related use, changes require an appropriate impact assessment, suitable validation, and a justifiable decision on usage. Consider linked workflows, data sources, and access rights. Verify the actually delivered target state. A technical publication does not alone confirm that the modified application remains suitable for the intended process.

How do we handle a personally tied connection following a personnel change?

Verify ownership, authentication, and relevant processing activities. Plan a controlled transition and confirm the technical workflow. A change of owner name alone does not establish a valid data linkage. Designate a representative and define an escalation pathway to ensure operations and required documentation are not dependent on a single individual.

Is a previously performed App test still valid following a flow change?

Only to the extent that the statement remains transferable. A modified processing step may affect data, status, rights, or repeatability. Assess this dependency and re-run appropriate end-to-end cases. A unchanged interface does not automatically confirm unchanged scientific efficacy.

Sources and review date

The sources provide the technical foundation. Tables, checklists, and examples are derived methodological tools, not verbatim regulatory requirements. Manufacturer information must be referenced to the specific version and configuration in use.

  1. European Commission: EU GMP Annex 11, Computerised Systems (2011)Current version as per EudraLex inventory as of 14.10.2026. Application, risks, operation, and data lifecycle; specific testing proposals are derived independently.
  2. Microsoft: Application lifecycle management with Power PlatformSupplier source for product functionalities. Availability and behavior depend on version, edition, license, and configuration; no general compliance certification.

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Verify the complete app workflow end-to-end.

From your app, the associated flows and data sources, we clarify the intended use, critical rules, and appropriate reference cases for the provided system.

  • 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 →