How AI can make Inbox Zero more attainable


Busy people receive hundreds of emails a day. Conversations that 20 years ago would have been done face-to-face have morphed into loosely coupled emails, instant messages and texts, PowerPoint presentations, Skype meetings and, yes, even a few phone calls.

And while electronic communications has made collaboration across global teams possible, it has not always improved overall communications. If anything, it has put the shackles on us all, leaving most people in a situation where they make a desperate attempt to prioritize the messages, always wondering if something important slipped between the cracks.

Here’s where artificial intelligence and the use of smart assistants can help solve some of our communications issues.

By applying machine learning and Bayesian logic, I can create a personalized, prioritized list of my emails. Interestingly enough, the technique is not new. We have been using these techniques to identify spam for 20 years. Today, we can easily create a score (rank) for every email – based on the user’s preferences – that learns what the user does and does not read.

A message sent from my wife just to me will always be high priority. Likewise, email from my bosses at work. A message from the group of people I regularly work with might arrive as a Priority 2, but a day later it might go up or down based on the content. Was I given an action? Did they ask a question? If yes, then the priority should increase the longer I don’t reply.

When I start a new project, the assistant will learn that I consistently open an email from James and respond quickly. The assistant will learn that this is an important relationship and automatically increase the ranking. Once the project finishes, the assistant will similarly notice that I no longer read/reply to emails from James and slowly the weighting will reduce.

Similarly, I might subscribe to a mailing list accidentally and read the first few and then as the backlog of email builds up the original emails become noise. Initially, the assistant will reduce the weighting of these e-newsletters, then start auto-filing, and then after a while the assistant will unsubscribe me completely.

Unlike spam, where users have to tell the system that an address is OK, the ranking of a message would change based on the content — the message’s sentiment combined with time, the number of recipients, the previous message, topic history, and even subsequent messages.

When a message arrives, it might rank as priority 5. Later though, another message might arrive on the same topic, but now the tone has changed, it has increased visibility. Maybe my boss or the head of a region is not on cc:, or maybe it has become more urgent, such that the rank is adjusted dynamically to increase the message’s visibility in my queue.

On the flip side, an email that’s a month old is probably no longer important. It can be filed/archived without risk. Under this scenario, Inbox Zero becomes a reality for everybody. Now, we can focus on what’s important and relevant to us. Instead of just presenting the information Last In, First Out, the Inbox helps structure the tasks, activities and communications we are bombarded with every day and brings some order and prioritization.

I can hear the screams about privacy already. Two points: First, the digital assistant can be deployed as a local agent. Second, how is this different than having a human assistant organize and coordinate an executive’s life? Time is money, so the potential for increased productivity and cost savings can win the day.

Have a look at more of Marc Wilkinson’s thoughts in Wired.

Marc Wilkinson

Marc Wilkinson was DXC Technology’s chief technology officer for Workplace & Mobility. He left DXC in October 2019.


How AI can make user support more proactive

Delivering proactive and predictive IT user support

AI and chatbots will revolutionize how we work


Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.