Header: AI Makes Super-Notes from a Federal Meeting.
##
Today, CMS held its professional advisory panel for pricing of new lab tests for 2027. The meeting was broadcast by Zoom.
Here's what AI can do.
1. I ran an audio auto record during the Zoom broadcast.
2. Otter.ai made an auto transcript (somewhat rough, not great formatting).
3. I gave two documents to Chat GPT. (1) The CMS PDF agenda for the meeting, 110 entries in a table. (2) The autotranscript.
I asked CMS to clean up and format the entire autotranscript. Add, break it into sections for each agenda item
For each agenda item, report from the PDF, the FACA PANEL item number, the agenda number from the June public comment meeting ALM, the code, and also the long descriptor. THEN print the transcript from that minute's discussion.
Chat GPT chugged away for about 15 minutes and then delivered the output (for the morning session) in about twenty minutes as a Word doc (as I'd requested)
Prompt at bottom.
##
Screen shots of the auto transcript, the agenda, and the output below. (Click to enlarge.)
##
##
Prompt
This is an autotranscript of the morning session of the summer Medicare pricing meeting for new lab tests. This is the session where a panel of experts comments on the pricing of each new code.
You have (1) the raw autotranscript and (2) the meeting agenda pdf.
Please edit the transcript for clarity, but preserve it as basically a transcript. For example, adding or deleting individual words or rearranging grammar is fine, but deleting one and more sentences would go too far.
Start each new code with a new section. Use the PDF agenda to provide the FACA Panel # in sequential order, the ALM Code #, the Current Code, and the Final Code, which is often TBD. Then insert the full long code descriptor, followed by the edited and cleaned transcript text.
Please export the result as a Word document, if possible. If not, choose another appropriate format.