TL;DR
First hackathon. 24 hours straight. Four people, one table, laptops open. We built "Project Sarah" -- an AI-powered B2B Sales Agent with Next.js, Supabase, and Claude. Scrapped the first idea after 2.5 hours of brainstorming. The submission video had no audio. Still one of the best days in a long time.
AI insights for decision-makers
Weekly. Practical. No spam.
Friday, 4 PM, House of Innovation
150+ people in the room. OpenClaw Hackathon Vienna. I barely know anyone here. Some came through the Entrepreneurial Leadership Program by Austrian Startups, others just showed up. We form a team of four, grab a table, power, WiFi. Let's go.
My expectation: meeting like-minded people. Maybe build something, maybe just talk. No agenda. I just wanted to experience what a hackathon actually feels like.
The First Idea? Scrapped.
2.5 hours of brainstorming. Going in circles. The idea is too big, too vague, too much at once. Then the moment that turned out to be the most important one: letting go. Scrapping everything, rethinking, focusing.
It wasn't the technology that mattered. It was the decision to drop an idea that wasn't working. That was the real breakthrough.
Project Sarah: B2B Sales Agent in One Night
What we built: "Project Sarah". An AI-powered B2B Sales Agent for the DACH market. Automated lead research via Apollo.io and Google Maps. AI-based scoring and qualification. The stack: Next.js, Supabase, Claude as the AI engine.
I wrote a lot of code, but spent just as much time on structure. Creating issues, setting up the architecture, distributing tasks. My experience with multiple parallel AI projects helped me bring patterns to the table that made execution significantly faster. When someone was stuck with Claude Code, I jumped in and helped.
By the end, everyone experienced what I experience daily: that flow when working with AI tools and suddenly building things in hours that would normally take weeks.
Grocery Store Snacks, Sushi at 11 PM, McDonald's at 3 AM
In between, everything you'd imagine: grocery store snacks, sushi at 11 PM when the sugar crashed. McDonald's at 3 AM. Squats in the hallway at 4 AM because backs were giving out.
The laughter from our room attracted other teams. Everyone talked to everyone, exchanged ideas, helped each other out. Nobody had planned to stay all night. But the energy in the room was such that nobody wanted to leave.
The Video Without Sound
Then morning. Deadline 11 AM for the submission video. We're fixing bugs until 10:45. Then: prepare everything, walk through it structured, three minutes max. We knew it had to be a one-shot take.
First attempt: again. Second attempt: nailed it. Everything ready to upload.
Then playing it back: no audio.
An explainer video. Without sound.
Down to the jury, asked for an extension. Granted. Back into a room, everything from scratch, this time with audio. Uploaded, submitted. Pure dopamine.
What I Took Away
Three things that stuck:
-
The most important moment is when you let go of an idea. Not when you start building. Scrapping the first concept gave us more than any single technical decision.
-
Structure beats talent at night. Creating issues, architecture before coding. Sounds boring, but at 4 AM nobody has an overview without a system.
-
The flow with AI tools is contagious. People who were skeptical before understood after one night why I work this way. Experiencing it is different from reading about it.
At 5 PM I collapsed into bed and slept until 6 AM. Foggy for two days after. But above all: I want more of this.
This was my first hackathon. I had no expectations. They were exceeded by far. Definitely not my last.
The project is open source: Project Sarah on GitHub.
Have an AI project in mind?
Let's analyze your potential together.



