To deploy an Azure Custom Vision model to a mobile device (Android), the model must be compatible with edge deployment. Standard domains, such as "Retail," run only in the cloud and do not support export.

1. Change the model domain: You must first update the project settings to use a Compact domain (e.g., "General (compact)"), which is optimized for real-time classification on mobile devices.
2. Retrain the model: Changing the domain invalidates the previous training iteration. You must retrain the model so the service can generate the lightweight weights required for the compact architecture.
3. Export the model: Once trained on a compact domain, the "Export" button becomes available, allowing you to download the model in formats like TensorFlow Lite for Android.

Microsoft Learn. (2023). Export your model for use with mobile devices - Azure AI services. Retrieved from https://learn.microsoft.com/en-us/azure/ai-services/custom-vision-service/export-your-model

Relevance: This official guide explicitly details the workflow: "To export your model, you must select a compact domain... After you select a compact domain, save the changes and train your model... The Export button appears."

Microsoft Learn. (2023). Select a domain for a Custom Vision project. Retrieved from https://learn.microsoft.com/en-us/azure/ai-services/custom-vision-service/select-domain

Relevance: Confirms that "Models generated by Compact domains can be exported to run locally," whereas standard domains (like Retail) cannot.