Related: letting your code fail assuming you will be notified and can fix the underlying data or code issue is FAR FAR better than a bunch of `if` statements to check for invalid states or work around them.
If you have designed your service layer and domain model appropriately, most routines can be a so-called "happy path" and exceptional cases, such as invalid data, can simply blow up and notify you via your exception handler (e.g. Bugsnag).