One way of adding data is to import existing NLU data in Markdown of JSON fornat.
You can also copy and paste or type a bunch of sentences in Training Data > Insert Many and select an intent to tag them all at once
You can also use Activity > Populate to add data.
If your model works decently, this can be a fast way to enhance your model: add a few examples, Rasa NLU will interpret them and display the annotated results in the New Utterances tab. Making corrections will be very fast, and you might find out that some examples are not that useful. Remember, you should only add more data when it increases the accuracy of your model.
Finally, you can add examples one by one from the NLU playground. Just click on the interpreted example to open the editor:
Then make your annotations and save.