Hai Dan Ives
That is a great question thanks for using QandA platform
i guess the speech Studio's "Custom Keyword" feature does not support retraining the model with additional sample data directly within the tool itself. for morre accuracy, gather more diverse data and consider training a new model using frameworks like TensorFlow. Transfer learning can also refine an existing model. Once improved, deploy the model to your preferred infrastructure. Speech Studio simplifies model creation.
If this helps kindly accept the answer thanks very much.