Runway

Inspiration

We took home a track win at a previous hackathon for resolving a huge cybersecurity problem for small and medium sized businesses.

We realized this was a space where we were not just inspired to do great work, but also was not being touched by recent waves of AI startups. We decided to think about the problems a grandma operating a small bakery or two sisters operating a hair salon would run into, and cashflow popped up in our heads. This gives rise to Runway.

What it does

Runway is an easy to use agentic dashboard and alert system connected to transaction and banking platforms.

Runway does not show you a lot of metrics you can't spend time learning and understanding.

It simply shows you the most pressing indications extracted from your finances. It rapidly identifies a cashflow problem from your finances, and shows you, e.g. that there is no way you're making payroll for your employees for next month. It then outlines things you can do immediately (follow up on huge catering invoices that have gone unpaid, but would greatly help resolve your issues making payroll). It then CALLS you via Eleven Labs with urgent alerts based on severe cashflow or customer problems you might be having

How we built it

We built Runway in a single Next.js 14 codebase and divided ownership by technical layer so each teammate shipped end-to-end: Abhimanyu built the Supabase data layer, migrations, and mock Stripe/banking APIs with seeded transaction data; Nicolas built Gemini-powered categorization plus the cash-flow forecast and runway logic; Arya built the dashboard experience with Next.js, Tailwind, shadcn, and Recharts; and Vidur implemented the alert scenario engine and Azure Communication Services SMS notifications. We used production-shaped mock financial endpoints to move fast in the hackathon window while preserving a real architecture, with a backend pipeline of connect → ingest → categorize → forecast → detect alerts → notify and Supabase as the system of record for businesses, transactions, alerts, and scenarios. ElevenLabs helps us call users to notify them of dire financial situations automatically. Auth0 was an excellent lightweight authentication framework that really well integrated with some of our phone calling functionality

Challenges we ran into

We had to ditch the idea of SMS alerts but the complexity of Twilio and Azure's regulatory requirements threw us off. ElevenLabs' voice agents was an excellent strategic pivot, not only is it a phenomenal project sponsor, but the variety of voices we could tailor to our usecase of alerts and urgency was amazing.

Accomplishments that we're proud of

We're proud of really diving deep into the problems of small and medium sized businesses and not resting on the laurels of previous successes. We spent a lot of time just focusing on how much value we can add to the life of a grandma who shouldn't have to focus on cashflow optimization and resolving the granular issues of her business. There's a lot of time we could spend talking about technical accomplishments and the agentic experience, but we want to highlight our method behind coming to this solid value proposition instead.

What we learned

There's a ton of problems to solve in every idea for every track. Distilling the problems down into a few big ideas that can be solved in the duration of such a hackathon is more key than having high quality developers on the team.

What's next for Runway

Now that we've found agentic solutions for cybersecurity and cashflow for small businesses at multiple hackathons, we will seek to add to our suite of hackathon products for SMB's via compliance and capital-seeking experiences.