It might be neat to apply machine learning to something like incoming messages (whether E-mail or text or otherwise), to automatically determine the real importance of a communication based on how you classify other messages.
After awhile, it should be possible to achieve minimal interruption, based on what you consider worthy of an interruption.
I know that I don’t like simple solutions like marking E-mail as “important”. In my experience, some people will always mark their stuff as “important” and their definition is not my definition.
I think Thoughtbot did something like this with 'FOMObot' - analysing whether you're likely to be interested in current conversation in a channel, rather than always being altered to the channel's activity.
Iirc it's open source on Github, I heard about it on 'the bike shed' podcast. (Which I'd recommend, for whatever it's worth.)
My understanding of what makecheck is talking about is a more personalized sort of classification, though.
I haven't used Spark, but it seems from what I see that this aspect of the product just classifies a few generic types of emails that everyone receives.
However; I don't think Spark learns to classifies emails that only I receive. Like I always instantly delete emails from Debby in marketing about upcoming customer engagement sessions.
If you could create your own category, and add new emails to that category and once you have a statistically relevant number, ML could start classifying that stuff.
On the other hand, who needs ML when you can just set up a filter to automatically delete emails from Debby with the word "Customer Engagement" in the title.
After awhile, it should be possible to achieve minimal interruption, based on what you consider worthy of an interruption.
I know that I don’t like simple solutions like marking E-mail as “important”. In my experience, some people will always mark their stuff as “important” and their definition is not my definition.