I've been working on an idea that I've found to be very useful for myself and I would like some feedback on. As an engineer working remotely at a large company I find myself in a lot of meetings. The main problem I have is that I struggle to remember the details of meetings after a day or two. I've tried to take notes but typing throughout the whole meeting is exhausting and I lose focus on the discussion. Oftentimes meetings such as a weekly project review are recurring and the time period is too long in between to remember a high level of detail about what happened last time.
Seeing as I'm an engineer I've decided to automate my note taking experience. I've been recording the audio of these meetings using a program I wrote and saving the transcripts for review before the next meeting or at the beginning of the next day. I plan on implementing some AI to summarize the text of the meeting and identify speakers in the audio.
I would think that this tool would be useful for pretty much anyone working remotely and is in a lot of meetings which is a very large audience. I envision building a moderately sophisticated app that spreads out these searchable meeting summaries on a timeline and adding some intelligence around summarizing groups of meetings into units like a project or a meeting series.
I've always heard it said to build something you yourself use so I thought this might be worth pursuing. I would start small but would scale up the complexity and value proposition over time.
I'd love any feedback or thoughts on the idea!
If the meeting doesn't end up with action items, nothing in the meeting was important really
Meetings minutes rarely get read
Usually the people trying to collect all the small details are the newer ppl that are still learning either the job, topics or ppl in a way
As much as it sounds like the right thing, I think the actual usage for it would be relatively light, it certainly not something everyone would use due many might say it's a good idea
I assume you can make a really easy MVP using existing services to transcribe and to summerise, don't reinvent the wheel for an MVP even if you won't get a feature like speaker identification