A matching TXT file can show a subtitle for an audio clip. Matching PNG, JPG, or WEBP files can give that clip an image, while one fallback image can cover the whole pack.
Start from the clip you already tested. Everything attached to it keeps the same base filename.
For hello_there.mp3, the subtitle becomes hello_there.txt.
Create a same-name TXT file with the line you want players to read. Check spelling and context as if it were the only clue to the impression.
Use an image only when it helps players recognise the prompt. Keep the asset clearly tied to the clip rather than adding unrelated art.
Add _pack_filler_image only when a shared Pack image makes more sense than individual pictures for most clips.
A different base filename will not attach the subtitle or image to the clip. File extensions alone do not create the association.
Ready to test
Quick answers
The filenames do most of the linking work, so small naming mistakes can hide otherwise valid assets.
Next step
Nested packs can organise a larger collection without changing the core matching-file rules.