InboxZero the Easy Way
My dirty little secret is that I have always wanted to become a software developer. The ability to build solutions from nothing is a dream come true, and when I was looking for pivots from clinical ABA work, I thought that software was the answer.
And while I am very happy I ended up in the marketing world, I think that there will come a time when I stop screwing around and commit to learning Python and JS and building my own solutions.
But, lately, I realized that I don’t have to wait. All I need is some AI, solid prompt engineering, and some low-code automation platforms, and I can do a hell of a lot. This simple email automation tool represents a big step for me. It was the first time I built something that helped my personal life, not my business or client life. Not a client project, not a way to get more leads, just a tool that makes my life easier.
As simple as it is, I suspect I will look back on this email automation tool as the build that started them all!
What was the problem
I have three email accounts that I use (technically 5, but two of them are forwarding accounts). One for personal, one for the book club, and one for the marketing agency. Because the marketing agency has a lot more customer-facing work, I typically am in my marketing agency inbox the majority of the day for important customer matters.
But, I also use the marketing agency email for most of my logins, which means it gets flooded with promotions as well… just like my other two emails.
Seriously, I have a problem with the number of subscriptions I have or have had in the past.
So, I was stuck in a situation where I had customers contacting me from all three sources, with financial documents and receipts coming in as well. I began to notice that there was a direct correlation between the number of emails that were unread in my inbox and the amount of stress I currently had. After all, each email was a signal for variability; it could either be reinforcing or punishing, and the uncertainty would make escape-maintained behavior much more valuable.
Did I just sneak in a casual reference to trauma-informed care and how unsignaled contingencies can increase the MO for escape behavior?
In the end, I needed a simple solution. I needed every incoming email to be sorted into different categories and archived from the inbox so that I could focus on the most valuable “to-do” related emails first, then filter to the less priority emails towards the end of the day, when I was out of creative steam.
What I did to solve it
I looked for a while online for a solution, but I found two different issues. Every system either could not handle multiple inboxes for a reasonable price, or they would not archive the email after labeling it. In fact, the only solution that I could find was hiring someone to help me do it manually. I went on Fiverr and hired a VA for $15 an hour to spend one hour per day categorizing each inbox, ultimately resulting in a somewhat clean inbox.
Unfortunately, I was now paying close to $300 a month for someone to just sort my inbox, much less reply, and I was getting increasingly uncomfortable having someone else in my inbox and aware of my password reset codes. So I started down the Google and YouTube rabbit hole to find another solution.
Eventually, I stumbled upon someone who was classifying text from PDF invoices and sorting them into different Google folders, and I realized I had found my solution. I had been building out custom AI lead gen and enrichment sequences for my marketing clients, so I had a decent amount of experience in the low-code automation space. I realized that if I could trigger an automation every time an email was received, I could then trigger an AI to consume the email and label it correctly, thus solving my problem.
Overall, the automation took me about 30 minutes to build, and about 90 minutes to test and perfect. But ultimately, 2 hours’ worth of work gives me an almost completely clean inbox 100% of the time, whether the email is received at 3 PM or 3 AM.
Here is how it works
- Every email triggers a webhook that sends the information to n8n (low-code software) with the content of the email.
- n8n passes the contents to an AI to read the email and sort it into 3–5 specific labels (depending on the inbox).
- Gmail receives the response from the AI, labels the email, and then archives it.
- I log into my email and glance at my labels, and I can tell I have received 20 promo emails (ignore), 3 notification emails (check later), and 2 lead requests (check now). 25 emails to 2 actions to take right now.
Conclusion
I thought I was going to have time during this post to go into how building this simple project was a big breakthrough for me as someone who considers themselves a product builder, but I won’t have time for that today. What I will say is that while the basics of this automation are very simple, if I were to ever take this outside of the “personal use case” and into an actual solution it would have to get much more complex.
From a more robust front end to bug fixes to UI/UX, this taught me that when it comes to software, solving the user problem is typically the easy part. The harder part is packaging that solution into a usable and scalable product.
Current Post
- Post Number: 7
- Words: 943
- Writing Time: 35 minutes
- Editing Time: 25 minutes
Overall
- Total Words: 6,633
- Total Writing Time: 222 minutes
- Total Editing Time: 134 minutes