Slyngshot: Mobile, AI-first redesign of an idea-to-pitch wizard
Mapping how founders, students, and schools used a web-based pitch wizard, then redesigning it mobile-first and AI-assisted — the app drafts, the founder decides.
Category: EdTech / Entrepreneurship
Status: Shipped on iOS and Android

Slyngshot is a platform from the Singleton Foundation that guides first-time founders and entrepreneurship students from a raw idea to a standardized pitch, no business jargon required. The Foundation hired our team to define and build the next generation: a mobile-first, AI-enhanced version of the original web wizard. I led discovery, mapping the existing user needs and architecture in Miro, which showed Slyngshot was serving more than solo founders — schools used it to teach, and organizers ran case competitions on it, so the next generation had to support shared, group-based use alongside the single-user wizard. From there I built low-fidelity prototypes in Miro showing how generative AI could help differently at each stage of a pitch: image generation for branding and identity, text generation for pitch content, and mobile-friendly versions of the existing funding-metric calculators. The interaction model that came out of that work was that the app drafts and the user decides — a first idea generates a set of assumptions, which in turn produce generative UI content choices the founder can edit, replace, or reject, so the founder stays the author of their own pitch.
We brought on an engineering team to build the resulting app, published to both iOS and Android. I owned the design work from discovery through handoff — leading the high-fidelity design and prototyping in Figma, building the mobile component library design and engineering shared, and negotiating trade-offs with the engineering lead when technical constraints put the intended experience at risk. The shipped release included an onboarding flow that captures a founder's first idea, a notebook-style home screen holding all of a user's ideas, shared spaces for classrooms and competitions, and an AI-assisted workflow — generated assumptions, content, and imagery — with edit controls on every output. This was early design work on generative UI and AI collaboration patterns, done before common conventions for either existed.