It looks like I found the solution.
I can do a case-insensitive text search and later in the next aggregation step I can do $regexFindAll to find all the occurrences for each found document and manipulate it as I want.
Thank you very much for your help.