Robotify

Inspiration Team BEAN was a green team with many problems in playing the game, understanding the strategy and scouting. 40% of the rookie teams abandoned because they did not have a mentor. So we decided to make one. What It Does AI App: 5 modes to learn Game Rules, Q&A, build strategies, get Rookie Checklist, and Smart Scouting. In real-time, Scout teams click 'Analyze', and receive a ranked alliance pick list that indicates which teams win competitions. How We Built It Create a full-fledged web app with React + Tailwind + Groq API in 8 hours. Five components, one AI function, useState data. System prompt baked using REBUILT game rules. Deployed to Vercel. Challenges 5 for 8 hours meant reducing scope. In order to prevent AI to hallucinate incorrect rules, they had to be extremely specific with the prompts they gave. Analysis logic for scouting was fairly complicated. Have some pre-loaded demo data for judges to view and show immediate results. Accomplishments Both 5 modes shipped and worked well. Scouting analysis addresses a problem that actually exists (2 hrs to 30 seconds). We had credibility because of Team BEAN's 60-8-0 record. Produces work at a MVP level in 8 hours. What We Learned Domain expertise matters. Good prompts are better than good models. Products that are “perfected” don't ship as MVPs. Lived experience is a source of credibility. There are 90K robotics teams in the market. We can scale this

Built With

• css
• html
• json
• react
• typescript
• vite