Entrepreneur-ing In Public: Hello World!

I'm teaching entrepreneurship by publicly building a startup, and you're invited to follow along.

If you’re reading this, you probably know I teach entrepreneurship at Duke University. You may also know, as part of my teaching efforts, I’ve decided to build a new startup in public. It’s an A.I.-powered email follow-up tool called Autopest. My goal is to take Autopest from $0 to $100k/year in revenue while publicly chronicling the entire journey. I call it “entrepreneur-ing in public,” and this is Entry #1 in that chronicle.

What is entrepreneur-ing in public?

Entrepreneur-ing in public is my personal twist on a niche phenomenon in the startup community called “building in public.” Building in public is basically the opposite of a stealth startup. It relies on radical transparency to promote a startup. Founders share everything — revenue metrics, growth numbers, marketing strategies, etcetera — in order to encourage the public to follow along with their startup journey (and hopefully buy the product along the way).

I love the concept of building in public, but it doesn’t quite fit my needs because it’s more of a marketing strategy than an educational tool. It’s like reality TV… real, but staged for profit.

In contrast, I’m mainly focused on teaching entrepreneurship. As a result, I’m not building in public; I’m entrepreneur-ing in public. Yes, I’ll share plenty of details about the company I’m building, but I’m mostly interested in sharing lessons about building companies and the difficult work of being an entrepreneur.

What is Autopest, and why’d you choose a stupid name?

In order to entrepreneur in public, I needed to build a startup, and Autopest was the closest thing to a startup I had lying around.

Autopest didn’t begin life as a startup. Instead, it began life a decade ago as a hacky piece of software I created for myself. Specifically, back when I was building startups (rather than teaching entrepreneurship,) I was sending hundreds of emails every week — to customers, employees, partners, investors, and so on — but people didn’t always respond. In those cases, I needed to send follow-up emails, and I often found myself either forgetting to follow up or wasting hours sending the same, basic message: “Hey… did you get the email I sent the other day?” I got so annoyed about how much time I was wasting sending follow-ups, I decided to automate the process by building a small web app to do it for me. It would check my email to see if people responded to the messages I’d sent. If they didn’t respond within a few days, the app would automatically send a follow-up. I called it Autopest.

Autopest worked incredibly well… for me.

I thought about productizing Autopest when I originally built it, but I couldn’t figure out how to make the automated follow-up emails dynamic. In other words, Autopest worked for my purposes, but, because the follow-up emails weren’t editable, it couldn’t work for everyone. Sure, I could have allowed people to create their own follow-ups, but doing so felt like over-complicating the product. I wanted it to be simple, and if people still had to write their own emails, what was the point?

That all changed ten years later when A.I.-powered language computational models like ChatGPT came along. Personally, when I originally saw ChatGPT, one of my first thoughts was about Autopest: What if I used the same technology to craft custom follow-up emails based on a person’s original message.

To be fair, even after recognizing I could use A.I. to power Autopest, I wouldn’t have done anything with it. After all, I was a “retired entrepreneur” who wasn’t building startups anymore. But then I decided I wanted to entrepreneur in public and… well… here we are. I dusted off my old Autopest codebase, plugged in the ChatGPT API, and — boom — Autopest 2.0 with custom email follow-ups powered by artificial intelligence!

Was it really that simple?

As with every seemingly brilliant startup idea, I quickly realized my “simple” plan was exponentially more complicated than I’d initially thought. In fact, let’s call this Lesson #1 of entrepreneur-ing in public:

In the startup world, everything takes at least 10x longer than you expect.

In my case, I thought I could make a few tweaks to my old Autopest codebase, maybe add a new “skin” on the frontend, and I’d be ready to launch. It was going to take me a week. Maybe two weeks… max!

Four months and lots of frustration later, I finally launched. Along the way, I built an entirely new codebase using a completely different programming language and database architecture. I also decided I needed a Gmail plugin to make the integration seamless, so I spent a couple weeks building that, too. And then, when I thought everything was finally ready, I ran into something entirely unexpected with the Google API that Autopest uses to send follow-up emails. Apparently, Google has a 10-week auditing process to use its API that can cost up to $50,000.

Imagine, for a moment, that you’re me…

You just spent months of your life and hundreds of hours building a micro-service startup just because you wanted to blog about it a couple times per week. However, the moment you finally think you’re ready to launch, you discover it’s going to take 10 more weeks and cost thousands of dollars.

I was pissed.

But, like any good entrepreneur, I persevered. Thanks to lots of helpful strangers on the Internet, I found a way to do complete the Google audit without spending any money. I even finished the process in six weeks rather than 10, which didn’t feel like a win at the time, but, in retrospect, could have been worse.

That brings me to here and this article…

What is Autopest’s current status?

As of this writing, Autopest is launched and operating in “public beta.” That’s my startup-y way of saying it’s working, but I’m still tinkering with things, so please cut me some slack. After all, I’m just one dude, and I’m managing the frontend, backend, database, APIs, Chrome extension, marketing, billing, and customer support for an entire company while also blogging about it and maintaining my full-time job of teaching entrepreneurship at Duke.

At the same time, I’d like to take a moment to pat myself on the back because, if I’m being honest, Autopest is one of the coolest apps I’ve ever built. To be clear, it’s not a billion dollar tech company, but that’s not what it’s supposed to be. It’s supposed to be the following two things:

First, Autopest is a micro-service. It does one, relatively small thing for a relatively small audience. But it does that one thing incredibly well. If you’re someone who sends lots of emails and doesn’t always get as many responses as you’d like (e.g. an entrepreneur, salesperson, account rep, manager, etc.), you should check out Autopest. It’ll make you’re life easier. Plus, unless you’re a power user, it’ll be free.

Second, and more importantly, Autopest is a public experiment in entrepreneurship. While it might turn into an enormous success, let’s be honest: It’s much more likely to be an embarrassing failure. Luckily, I’m OK with failure so long as my embarrassment helps other people be better entrepreneurs.

-Aaron