How I used deep learning to get better at managing Emails 📩

Image for post
Image for post
Photo by Diana Akhmetianova on Unsplash

Like most people, I get a ton of emails at work every day. And I am not great at managing through them. Forget being great, I am not even good.

It doesn't help that my day job is not about sitting behind a computer all day. Instead, I am on the road meeting customers and hustling with the sales team to secure orders. And when I am in the office, my goal is to be distraction-free and try to get at least 3 things done before the day is over.

So what happens to all the emails I get every day?

99% of the time, I read them on my phone, to see if something EXTREMELY pressing needs my attention. Note that I am mostly ‘skimming’ the emails — making a mental note about the ones that I need to reply to.

I will check emails en route to a customer meeting, thinking “I need to respond to this after the meeting” — but then I will forget as other things take priority. It doesn’t help that as I skim through the emails, my phone keeps marking them as ‘Read’.

But then here is when things REALLY start going south. Before I can read/respond to these emails, the day is over — and another truckload of emails arrive. Old emails seeking a response keep getting buried under new ones.

I wanted to fix this.

I tried a few productivity apps with bells and whistles — but they ended up making simple things more complicated. I didn’t want to change the way I work because other than managing emails, I am quite productive. For instance — I didn't want to adopt popular productivity advice such as “reply to emails as you get them” or “just take an hour on the weekends to read emails and clear out your inbox”….. because well that's not how I work.

Instead, I needed something much simpler.

I needed someone/something that could do just the following 3 functions

  1. Read all the emails I receive each day, and mark the ones that are seeking a response
  2. Add these emails to my to-do list
  3. Remind me periodically, so I don’t forget to respond to these emails

I needed an Assistant 😎

But I don't have one. So I decided to build an assistant that would do the 3 things I mention above. And functionality #1 was going to be the trickiest.

Emails come in all forms and sizes and different people have different writing styles. So, how could one write software that would reliably ascertain whether or not the email sender requires a response?

Deep Learning

I had no prior background in deep learning artificial intelligence, but in 2019 I came across Jeremy Howard’s brilliant MOOC called Practical Deep Learning for Coders and it got me hooked.

I learned the principles of deep learning, how we could ‘train’ algorithms that can then do pretty much anything. Trivial things like looking at lots of cats and dog pictures and telling which is which. But also complicated systems such as movie recommendation engines based on ratings and one’s preferences and much much more.

I wanted to learn more and the best way to learn deep learning was to actually build something that uses deep learning. I knew the exact thing I wanted to build.

I decided to use deep learning principles to solve my email management problem — and built an AI assistant for Outlook for my work emails. This assistant looks at my emails and detects if an email contains a task. As I built the assistant, not only did I learn a ton not only about powerful deep learning techniques, but also the very real implementation challenges.

***To get detailed step-by-step posts on how I built the AI assistant — just subscribe to my newsletter below!

In any case, now I had an AI assistant that could Read all the emails I receive each day, and mark the ones that are seeking a response. Functionality #1 was accomplished. 💥💥💥💥💥

To achieve Functionality #2 & #3, I integrated the AI Assistant with a task management app that I built — Taskwillow.

Taskwillow is a very simple to-do app that lets you create, organize and also share tasks with others. It also notifies you periodically on the ‘to-dos’ that you’ve recorded or others have shared with you. Our AI assistant for Outlook automatically adds any emails (that contain tasks) to your Taskwillow list.

Image for post
Image for post

Sidebar#2: Taskwillow is FREE — but the AI Assistant functionality sits behind a paywall because I incur monthly expenses on the hosting infrastructure 😐😐.

I have now been using the AI assistant for a month, and it has really helped me to stay on top of my email game.

  • I don't prioritize reading emails with no calls to action or no implied tasks. skimming these emails is enough
  • I don't NEED to respond to emails right away because I know I won't forget to respond later.
  • And most importantly, older (unread & read ) emails that are awaiting my response are not getting buried -the assistant identifies them and sends me these reminders!
Image for post
Image for post

Thanks for reading!!

Watch this space as there is more to come but meanwhile, if you have a feature request for Taskwillow, just leave a comment below or reach out to me on twitter!

Written by

Building Taskwillow. Notes about building things and selling them. Sometimes other stuff too. I am on Twitter @chakrvyuh.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store