We are consistently missing the last part of the transcript. The audio recording provides full call, yet the transcripts misses the last part of the audio. Here's an example: Call ID - 1f19c657-4fd4-499c-bd8b-3cb77c179185 Timestamp (missing transcript from this point) - 1:39

Add 3–5 seconds of silence or a short filler phrase at the end of the call to prevent premature endpoint detection..