How I used AI to become an inbox zero person

You’ve probably seen the articles before about people who had tens of thousands of emails and magically took it down to zero in just a couple of days with the help of various tools, methods, or just being plain old fed up with having too many emails and sitting down to do the task manually. As someone who gets loads of emails for work, it’s an interesting concept that I wanted to try myself, and I was able to get some moderate success out of it. 

Before anyone’s hopes get too high, the whole inbox zero concept doesn’t work the way you think it does. Email is a fact of life for many jobs, and keeping it at zero at all times is a Sisyphean effort. You can get your inbox to zero, but it’ll fill back up again. Even the original inventor of the inbox zero concept, Merlin Mann, says that most people get it wrong, having taken his advice from years ago way too literally, and that the point isn’t to have no emails in your inbox at all. 

Instead, the point is to sift through fluff and bloat to get to what’s really important, and that’s not only much easier, but the various tools available today make it really easy to do. Below, I’ll describe the tool I used for it, and how I configured it to help me sift through the nonsense. 

My AI-powered inbox zero method

There are a variety of AI email assistants available on the web, but the one that seemed to work best for me was Zero Email. The primary reason for this is that individual accounts can use it for free, and while I do have multiple email addresses, I only have one for my job. I have gone nearly 30 years without paying for email, and I don’t intend to start anytime soon.

What I like about this tool is that it’s a nice mix of traditional email with just a touch of AI there to help you sort through things. The free version includes an AI bot that you can ask for help 20 times per day, and you can also edit the prompt to change how the bot acts around you. The UI shows your inbox, and you can manage your emails and talk to the AI bot from one interface. 

The value in this approach is that the bot can perform tasks faster than I could through Gmail’s native interface while still giving me nearly full control. I did consider more automated approaches like Zapier, Inbox Zero, and a few others. However, Zapier has quite a learning curve for something that I only needed long enough to help me organize things, and solutions like Inbox Zero cost money out of the gate.

Trimming the fat

The first thing I needed to do was reduce the number of emails I received. This would prove to be a pretty simple task, although it did take me some time. In the old days, I would’ve used an unsubscribe service like Unroll.me, but the whole idea kind of went out of vogue when Unroll.me was found to be selling user data to Uber. Instead, I used two native Gmail features and a little creativity. 

I started by creating a label within Gmail called “Junk E-mail.” The name is arbitrary, but the goal was to funnel emails into this label for me to manually sort later. Then, I created a filter in Gmail that would identify any email that had the word “unsubscribe” in it, mark them as read immediately, and ship them off to the Junk E-Mail label for me to sort through later. 

This works because all mailing list emails are required by law to have a clearly listed link to unsubscribe from emails. So, with these two simple things, almost every mailing list email was sent to one folder where I could manually inspect and unsubscribe from email lists. The only ones that squeaked through are ones that used an unsubscribe graphic instead of a word, but they were easy enough to weed out manually. 

I did this step first and let it percolate for a while before I did anything else. The results were immediate and effective. As a tech writer, I have signed up for hundreds of services over the years so that I could talk about them in articles, so the first few days of this was harrowing. Eventually, the emails slowed down and eventually became manageable, and then finally, a non-issue. Email lists I wanted to stick with were moved back to the Inbox for future processing.

Preparing the inbox

The next step was email preparation. Most of my emails fall into a handful of categories, and so preparing the inbox was an essential part of the plan. This part is doable manually, but AI can do it much faster. 

So, my first task was getting rid of every label I didn’t want or use. This was easy enough. I asked the AI to delete them, and it did. I then came up with a list of labels I believed would help me get through my emails faster, like Press Emails, a label for each of my freelance clients, and generic ones like Shopping. Another prompt to the AI, and the labels were immediately created. 

AI handled this entire process and it only took a couple of minutes. The longest part was figuring out what labels I wanted to make and which ones I wanted to delete. From there, the AI handled the rest.

Putting the emails where they belong

The final step was by far the hardest, which made me eternally grateful that the AI could do it all for me. This took a few days because of Zero Email’s daily limits, but the process was pretty much cut and dry. 

I started by asking the AI to create filters to put various contacts into various labels automatically. This process would’ve taken many hours to do myself as a function of having to bounce between browser tabs and Gmail’s filter’s UI. Instead, all I had to do was ask the bot to do it for me. This includes things like sorting Amazon emails into the Shopping label and non-work emails into a Personal label. 

The only part that was difficult to figure out initially was press emails. I get these from random companies all the time, and not all of them are in my contacts list. For this problem, the solution I came up with was similar to the one I used for junk emails above. I had the AI create a filter to sort any email that includes common press email words and phrases like “announce,” “launch,” and “embargo.” This filter required manual tuning over the course of a couple of weeks as more phrases and words were identified, but it has mostly worked well. 

Once this was done, the number of emails in my inbox without a label very nearly dropped to zero. The benefit here is that I could click through my labels to view the emails I wanted to deal with at the moment, and the rest could be largely ignored until I checked the label later. 

The benefits are palpable, and clicking into a label to see a handful of emails all from the same senders or regarding the same general topics was immediately less stressful than sorting through a completely disorganized inbox.

Clean up

Aside from some fine tuning, which I was perfectly capable of doing manually, the process was completely done and I was satisfied with my results after a couple of days of testing with everything in place. The last step was cleanup. 

To start, I had the AI go through every email in my entire Gmail account and label them with the above rules applied, just to make sure everything got sorted appropriately. The AI did struggle with this one a little bit but eventually did what I asked it to do. Everything was now where I needed it to be.

This concluded my business with Zero Email, and so the next step was removing it from my account. I get emails with non-disclosure agreements and embargoed information all the time, and while Zero Email’s privacy policy doesn’t raise any immediate red flags, this is information that people would be displeased to have leaked, and so, I try to keep the number of services connected to my Gmail at a minimum, even if the company promises it’s not reading my data. 

In any case, once everything was disconnected and the emails were all cleaned up, my new system was in place and ready to go.

Asking AI for help when needed

Gmail and Outlook both have AI built-in already, Gmail with Gemini and Outlook with Copilot. These AIs aren’t great at managing emails (yet). However, they are adept at helping you write them. You may have noticed to this point that I haven’t talked much about constructing emails