FishBowl

our 3 investors
example tony speaking (Text to speech)
Example Victor speaking (TTS)
Example Nana speaking (TTS)
Example 2 judges are OUT!
Mission failed screen.
TONY GIVING AN OFFER! (RARE)

Inspiration

Practicing a Shark Tank-style pitch is tough. It is high pressure, unpredictable, and hard to simulate without real people pushing back on your idea. We wanted to build something that lets founders rehearse quickly, repeat the process as many times as they want, and get blunt, personality-driven feedback from a panel that actually feels challenging.

What it does

Users submit a business pitch, and the app first checks whether the idea is clear and complete enough to move forward. It then pulls in a quick market snapshot and runs the pitch through three different AI Sharks, each with their own personality and style of questioning. From there, the user goes through multiple rounds of Q&A and negotiation that lead to deal-like outcomes.

How we built it

We built FishBowl with Next.js, React, TypeScript, Tailwind CSS, and Framer Motion. On the backend side, our server routes handle pitch validation and each Shark response using OpenAI, while Perplexity is used for automatic market lookups.

Challenges we ran into

One of the biggest challenges was keeping all three Sharks consistent in their own personalities while also making the overall conversation feel coherent. We also had to deal with getting structured outputs back reliably, handling timeouts and API failures, and adjusting parts of the implementation as the project evolved during development.

Accomplishments that we're proud of

We were able to build a full end-to-end experience where a user can go from pitch submission to validation and a multi-round Shark session in one flow. We are also proud of how distinct each Shark feels and how polished the overall user experience turned out, especially without relying on a heavy backend or database.

What we learned

A big takeaway for us was that building a good AI app is not just about the model itself. Prompt design, state management, and orchestration matter just as much. We also learned how important it is to keep track of the gap between the product vision and what is actually implemented, especially in a fast-paced hackathon setting.