The AI Assistant Needs to Manage Email Inboxes BetterThe AI Assistant Needs to Manage Email Inboxes Better
We give it enough data -- our email application should be learning more effectively from our reading and deletion patterns.
As my team and I sift through the transcripts and AI-generated summaries of our interviews and sessions at Enterprise Connect 2025, we've noticed a lot of folks talking about how generative AI copilots and digital assistants are boosting daily productivity. Meeting recaps and auto-generated action items are dominating, but you know what we're not seeing a lot of people tackle in the copilot space? Email management -- how to sort messages by their alignment with meeting-generated action items, or how to sift through messages based on correspondent history, or how to prioritize items.
Email is a time sink. Every time someone looks at their inbox, they have to scan the subject lines and senders, then decide, "Am I deleting this message or reading it?" And then once they commit to reading, the next question is, "Do I need to act on what's in this email? Right now? How long will it take? Is this going to derail my to-do list? Or add to it?" Even those of us who try to practice Inbox Zero with its five ways to handle a specific message -- delete it, delegate it to someone else via forwarding it, respond to it, defer acting on it until you have more time, or do whatever the email requests if the ask will take little time -- can easily get buried under the ceaseless barrage of messages, or look up after deleting 40 messages and handling another 20 "this will take only a minute" responses and realize 90 minutes have flown.
This perception of email overwhelm is not purely subjective. A late 2023 Microsoft Research study found that managing email chewed up over 20% of frontline workers' weekly working hours; the company discreetly noted, "for knowledge workers, who rely even more on digital communication, the share of the week taken up by emails and meetings is even greater."
Fast forward a year, and Microsoft reported, "Email overload persists—85% of emails are read in under 15 seconds, and the typical person has to read about four emails for every one they send."
Here's a great way to reduce that email overload: Make filters better. I'm not even asking for AI-powered assistants. I'm asking that the baseline automation behind a filter -- the if-this-then-that trigger -- get better. I'm asking that email app designers tweak the intelligence that allows for filters and actually use the data gleaned from countless messages to cut through the inbox clutter.
Here's an example: Every day, purported human beings from a random email address send me stilted messages asking about paid blog post placement on No Jitter. Today's example: "We are interested in working with websites like yours. We would like to publish guest posts about gambling and betting." While committing to specific communications platforms can feel like taking a big gamble, the betting industry is not one No Jitter covers. Yet here is the pitch for paid blog post placement -- nestled between similar pleas to take someone's money so they can run posts on No Jitter about CBD gummies, children's books and awareness-raising efforts for assorted chronic conditions.
None of these messages ever go to my spam folder or my non-focused inbox. And after every one of these messages, I immediately block the sender. But why must I? Why doesn't the AI presumably optimizing my inbox not look at a dataset of literally thousands of messages, scan the deleted ones and notice messages with the words "publish guest posts" are always deleted and the recipients blocked, then use that pattern to set up a rule for future emails so I don't even have to see those messages? Why am I still doing this pattern-recognition-and-response work with my human brain?
Were I a less scrupulous person, I'd consider collecting a fee from all those paid-placement correspondents and fund the development of the email tool that would improve my email client's capacity to notice what I don't ever respond to. Instead, I'll just ask when the AI assistant that can do that will be widely available -- and if it can handle all the emails that will inevitably be coming from the AI assistants meant to place all those gambling blog posts.
Read more about:
ProductivityAbout the Author
You May Also Like
