Most efficient way to update speaker names and refresh summary

General
1

I would like to know if someone has an efficient way to update speaker name in transcripts and regenerate the summary?

today, it is really messy. The speaker recognition part is 5/10 in terms of correctness, at best. I’d rather it just showed me speaker 0, 1, 2, 3 - I tag names once, it recognizes rest in the transcript and it generated a clean summary.

when i try to do this in the app, my iPhone literally heats up, the app seems to start to hang and i give up. And i have a meeting notes with 9 speakers when there were four of us in the meeting.

would appreciate any tips here from the community or team on best/easiest way to do this.

1 Like
2

Came here to ask this same question.

A great workflow would be, once I add the names in the transcript to replace “Speaker 0”, “Speaker 1”, etc.. it just does a find/replace in the summaries without regenerating.

4 Likes
3

I used the agent and asked it to change Speaker 0 to “Jane” and Speaker 1 to “My Name”. It went through and fixed it, at least it did for the summary.

4 Likes
4

So because we change the speaker in existing transcripts will it remember every recording afterward and apply that speaker? Meaning it should recognize the person on the recording and apply the speaker label

5

Yes, that is the idea at least. I know they are working on a few “work arounds” to help with this as the current AI tech is limiting this for us.

1 Like
6

I just got mine and have tried this several times and it doesn’t work unfortunately.

7

Hey @rjb just trying to understand the issue you are having. As mentioned in previous posts speaker splitting is something that is pretty nascent in the industry and we currently have active research running to improve our experience.

8

I’ve only tried a couple times but it seems like if I update speaker names in a transcript and later export that transcript, the export reverts to Soeaker0 etc. Is that intentional, and is it something we can change or work around?

1 Like
9

Hey @GarySlinger is this export on the app? If so we are checking this! Thanks for reporting

10

Hey @GarySlinger I am looking into this issue, will give an update on this, once it’s resolved, thanks for reporting

1 Like
11

Update on the Speaker Replacement Issue on Export:

We found the issue and fixed it, should be live on your app in sometime, thanks for taking time and reporting this issue @GarySlinger

1 Like
12

Yes - when I export, it doesn’t hold me edits. I haven’t checked yet whether it includes other edits in the text, but it definitely doesn’t include speaker name edits.

13

Whoops - working in sequence. Just saw this. Nice one. Will test shortly. Thanks.

14

OK so… Just tested. No joy. Tried it from a fresh started app, then regenerated the conversation - the regenerated conversation remembered the assigned speaker names, which was nice, but the export still has Speaker 0, Speaker 1. I checked for an app update in case it needs that - I’m at 0.5.23 if that helps.