The idea was fine. The execution was fine. The presentation was fine. We walked out confused.
Then we won three in a row: Prajukti 2026 (1st), Codewar 7.0 (2nd), NE Graduate Congress AI & Innovation (1st). Same team — me, Veeshal, Aakash, Aishik. Same general approach to building. One thing changed.
We stopped trying to impress the judges and started trying to make them feel something.
Making judges feel something
That sounds vague. Let me make it concrete.
In our losing attempt, our demo showed the product working. Features, flows, edge cases. We were proud of the technical work, and we made sure everyone knew how much technical work there was.
In Prajukti, our demo showed a problem being solved for a real person. We opened with a 30-second story about a small business owner in Assam who couldn't automate her workflow because every tool assumed she spoke English fluently. Then CrewSpace solved it, live, in her language.
The judges didn't ask about our tech stack. They asked about her.
The other things that actually mattered
Constraints breed creativity. Hackathons have time limits. Don't fight them — use them as a forcing function to decide what your product actually is. CrewSpace during the hackathon was one thing: regional language AI agents for non-technical users. Not a platform. Not an ecosystem. One thing, done well, in 36 hours.
Presentation is product. Your slides, your demo flow, your speaking rhythm — these aren't wrapping paper around the real work. They are the work, as far as the judges are concerned. We spent the last four hours of every hackathon on the presentation, not the code.
Know which teammate does what in the room. Veeshal handles the demo. Aakash handles technical Q&A. Aishik handles the business case. I handle the story and the narrative thread. We don't step on each other because we've decided in advance who owns what.
What I'd tell our losing selves
You built a product. Build an argument next time. The product is the evidence. The argument is what wins.