We recently purchased Macrium Reflect Deployment Kits and have successfully created our first Windows image. I understand that adding drivers to the USB boot or Rescue media only makes them available to the WinPE environment, not the Windows installation being restored. Is there a built-in Macrium feature for injecting the correct driver pack into the restored Windows OS based on the target computer model? I'd like to maintain one hardware-independent image and apply model-specific drivers during deployment. If Macrium cannot do this directly, would the usual approach be to run DISM or PnPUtil from a customized WinPE environment after the image has been restored?
3 Answers
For a more repeatable process, keep a minimally configured reference image or virtual machine template, then create and customize a copy with the required software, updates, and drivers. Run Sysprep when appropriate, capture the result, and deploy it. If you need one image across multiple hardware models, apply the model-specific drivers afterward through your deployment scripts rather than baking every driver set into the base image.
DISM can add drivers to the restored Windows installation while it is still offline. A custom WinPE workflow can identify or select the appropriate driver pack, then use DISM with the mounted Windows volume as the target. That keeps the base image hardware-neutral and applies the drivers before the system boots for the first time.
That was the approach I was considering—customizing the boot environment so it can inject the appropriate driver pack after the restore finishes.
Macrium does not appear to provide a built-in feature for injecting model-specific drivers directly into the cloned Windows installation. You may need to handle it externally with DISM or another deployment workflow. Tools such as OSDCloud can also deploy a generalized Windows image and apply drivers during deployment.
I’m already using a generalized image and Sysprep; the remaining piece is finding the cleanest way to apply different driver packs for each model during deployment.