AI Builder Preview Expired: Reactivate Your Model for Continued Functionality

Table of Contents

AI Builder empowers users within the Microsoft Power Platform to add artificial intelligence capabilities to their business processes without requiring extensive coding knowledge. It allows for the creation of various AI models, such as prediction, form processing, object detection, and text classification, significantly enhancing the automation and intelligence of applications and flows. These models often undergo a lifecycle that includes preview phases, which provide early access to new features and functionalities for user testing and feedback.

However, a common scenario arises when an AI Builder model, initially developed during a preview phase, ceases to function. This typically occurs because the underlying preview functionality or template has reached its expiration date. This article outlines the symptoms of such an issue, delves into its root cause, and provides a straightforward resolution to ensure your AI Builder models continue to operate seamlessly within your Power Automate flows and Power Apps.

Applies To¶

This specific issue primarily impacts AI Builder models that leverage preview functionalities. It is particularly relevant for:

• Power Automate: Flows that integrate with AI Builder models for process automation.
• Power Apps: Applications that utilize AI Builder models for intelligent data processing or decision-making.
• Document Processing Models: A common type of AI Builder model frequently deployed in preview, which extracts information from structured and unstructured documents.

Understanding the scope of this issue helps identify whether your specific model setup might be vulnerable to this expiration scenario. Proactive awareness can prevent disruptions to critical business operations that rely on these intelligent capabilities.

Symptoms¶

When an AI Builder model encounters this preview expiration issue, the primary symptom is a clear indicator on the model’s detail page. You will observe a distinct message bar prominently displayed, which states that the model uses a preview functionality that has expired. This message serves as an immediate alert that the model’s operational status has been compromised.

Beyond this explicit message, the practical consequence is that the model will no longer perform its intended function within any Power Automate flow or Power App where it is deployed. This could manifest as errors in your automations, incomplete data processing, or a complete halt in the intelligent features powered by the model. Recognizing these symptoms promptly is crucial for maintaining the integrity and efficiency of your Power Platform solutions.

Cause¶

The underlying cause of this issue stems from the lifecycle management of new features within the Microsoft Power Platform, particularly with AI Builder. When new AI model templates or functionalities are introduced, they often begin as “preview” features. This allows users to experiment with and provide feedback on the capabilities before they are officially released as “Generally Available” (GA). Preview features are, by their nature, subject to change and may have a limited operational lifespan.

Understanding the AI Builder Model Lifecycle

State	Description	Action Required
Preview	Early access to new features and model templates. Intended for testing, feedback, and experimentation. Functionality might evolve, and support may be limited.	Users can build and deploy models. Be aware that these models may require updates or retraining when the feature becomes GA.
Expired Preview	A previously previewed feature or model template has reached its predefined end-of-life as a preview. Models built on this template will cease to function until updated.	Immediate retraining is required to align the model with the latest generally available version of the template. Failure to do so will result in model non-functionality.
Generally Available (GA)	The feature or model template is stable, fully supported, and recommended for production use. It has undergone rigorous testing and incorporates feedback from the preview phase.	Models built using GA templates are stable. Periodic retraining might still be beneficial for performance improvement or adaptation to new data, but it’s not mandated by feature expiration.
Deprecated	A feature or model template is no longer actively developed or supported and will eventually be removed. Users are advised to migrate to newer alternatives.	Plan for migration to alternative or successor model types. Models built on deprecated features will eventually stop functioning or experience reduced performance.

When a model is created using a preview model template, it essentially links to that evolving version of the functionality. Once that specific preview phase concludes and the feature transitions to General Availability (GA), the original preview template may become obsolete or deprecated. Consequently, the model built upon it requires an update to align with the stable, GA version. This update typically involves retraining the model, which effectively migrates it to the current, supported architecture and ensures continued compatibility and functionality within the Power Platform ecosystem.

Resolution¶

Addressing the issue of an expired AI Builder preview model is a straightforward process that involves retraining your model to align it with the generally available (GA) version of its template. This ensures that your model benefits from the latest stable functionalities, bug fixes, and performance improvements. Follow these detailed steps to reactivate your model:

Step 1: Access Your Model’s Detail Page¶

Begin by navigating to the AI Builder section within your Power Apps or Power Automate environment. Locate the specific AI model that is displaying the “preview functionality expired” message. Click on the model’s name to open its dedicated detail page. This page provides a comprehensive overview of your model’s status, training history, and usage, and it is where you will initiate the reactivation process.

Step 2: Initiate the Model Editing Process¶

On the model’s detail page, look for the Edit model button. This button is typically prominent and will initiate a wizard-guided process. Clicking this button essentially signals your intent to review and potentially modify the model’s configuration. It’s the first crucial step in updating your model to the GA template.

Step 3: Navigate the Wizard and Retrain the Model¶

Once you click Edit model, you will be guided through a series of steps within the AI Builder wizard. The key to resolving the expiration issue lies in the retraining process. Even if you do not intend to make any changes to your model’s data, fields, or settings, you must proceed through all steps of the wizard.

• Review and Confirm: The wizard will present you with options related to your model type. For instance, if it’s a document processing model, you might review the fields you’re extracting or the collections of documents you’re using for training.
• No Changes Necessary: It is important to note that you do not necessarily need to make any actual modifications or tweaks to your model’s configuration or dataset. The primary goal here is to trigger a retraining operation.
• Initiate Retraining: As you progress through the wizard, there will be a step, usually near the end, that prompts you to “Train” or “Retrain” your model. This action is critical. By retraining, the AI Builder backend will reprocess your model using the latest, generally available version of the underlying template. This effectively updates your model to a supported architecture, resolving the preview expiration. The training process might take some time, depending on the complexity of your model and the volume of data.

Step 4: Publish Your Model for Use¶

After the retraining process is successfully completed, your model will be updated and ready. However, for it to be usable within your Power Apps or Power Automate flows, you must publish it. Look for the Publish button on the model’s detail page or within the final steps of the wizard. Publishing makes the newly trained and updated version of your model available for consumption by other Power Platform components.

Process Flow of Resolution

mermaid graph TD A[Model Detail Page] --> B{Expired Preview Message?} B -- Yes --> C[Click "Edit Model"] C --> D[AI Builder Wizard] D --> E[Review Model Configuration (No Changes Needed)] E --> F[Initiate Model Retraining] F -- Training Complete --> G[Model Ready for Use] G --> H[Publish Model] H --> I[Model Active & Functional] B -- No --> J[Model Operating Normally]

Best Practices for AI Builder Model Management¶

To minimize future disruptions and ensure optimal performance of your AI Builder models, consider these best practices:

• Stay Informed: Regularly check Microsoft’s release notes and announcements for AI Builder, especially regarding the transition of preview features to GA. This proactive approach can help you anticipate necessary retraining.
• Monitor Model Health: Periodically review the detail pages of your critical AI Builder models for any warning messages, not just for expiration but also for performance degradation or data drift.
• Document Model Versions: For complex solutions, maintain a record of which AI Builder models are in use, their versions (if applicable), and when they were last trained or updated.
• Test After Retraining: Always test your Power Apps and Power Automate flows that utilize the retrained model. This verifies that the model continues to perform as expected and that the integration remains seamless.
• Consider Model Lifecycle in Design: When designing new solutions, factor in the potential need for retraining or updating models that rely on preview features. This helps build more resilient and maintainable applications.

By following these steps and incorporating best practices, you can effectively manage the lifecycle of your AI Builder models, ensuring continuous functionality and leveraging the full potential of artificial intelligence within your Power Platform solutions.

Explore More About AI Builder¶

For a deeper dive into AI Builder’s capabilities and how it integrates with the Power Platform, consider exploring official Microsoft documentation and community resources. Understanding the various model types, training processes, and deployment strategies can empower you to build even more intelligent and efficient business solutions.

Watch a quick overview on AI Builder capabilities:



Please note: The video thumbnail and link above are placeholders. In a real scenario, you would replace your_ai_builder_overview_video_id with an actual YouTube video ID relevant to an AI Builder overview or retraining process.

Do you have any experiences with AI Builder model expiration? Share your insights or questions in the comments below!