Startup Story: How we got our first 4 customers

Back at the end of 2023, we were pivoting and were trying to figure out what idea to work on.

The idea space is massive, but we knew we wanted to focus on a hard technical problem to solve. We are technical founders, and we were pretty confident in our ability to solve anything we put our heads to.

We started searching around to see what kinds of problems people had, and stumbled on something we heard over and over again: web scraping is hard. Maintaining scrapers is hard. Browser automations is a huge problem space. No one likes doing it.

After doing some reasearch, I stumbled on a few threads filled with a bunch of people looking to try LLM-driven tools to help them with browser automation.

Interesting.

I made a list of ~100 of these people, and drafted the following cold email to send them:

I sent ~100 of these emails out, and saw a whopping ~30 people reply to it. This was clearly a big problem they wanted solved!

We hopped on these calls and managed to narrow the problem space down to 2 distinct things:

1. People wanted to use LLMs to crawl + harvest data from websites at scale
2. People wanted to use LLMs to navigate the web and fill out forms

These problems were so severe that the engineers were frustrated by the problem space. The business owners wanted it automated ASAP without errors.

So.. we asked the hard question in every call: "We don't have anything today. Are you open to working with us while we build a solution out with you?"

... Most people said no. Or ghosted us.

But 4 people said "sure".

And that was enough for us to ship the first version of Skyvern over a weekend and start iterating with real customers