AI Startup Coach
I built Underdog's AI Startup Coach for founders with an idea and no clear route forward. It helps them examine the idea and work out a next step they can return to.
- Team
- — Founder-led product
- Industry
- — EdTech, AI, Entrepreneurship
- Website
- — underdoginnovationinc.com
Tools & technologies
The Problem
Aspiring founders had plenty of advice and no clear next step.
The approach
How the work took shape
Learn about the business first
A blank chat box told the coach too little. I started with a business profile covering the founder's stage, goals, resources, and challenges. That context sat alongside startup guidance, practitioner knowledge, and academic research, giving each conversation somewhere specific to begin.
Leave with something usable
I built two ways to carry a conversation forward. Business Concept Evaluation lays out an idea's strengths, risks, and open questions. Personalized Pathways turns a goal into saved milestones, tasks, and timelines, so the founder can come back to a plan rather than start another chat.
Take it into a native app
I built the iPhone app in SwiftUI, with OpenAI and Gemini services behind the coaching workflows. Firebase stored the user's context and plans, while StoreKit, notifications, and access controls supported the App Store product. Separate services made it easier to change one AI workflow without rebuilding the rest of the app.
The hard parts
The constraint
Confident AI advice could sound more certain than the evidence allowed.
The decision
I structured the inputs and evaluation criteria, presenting the output as material for the founder to weigh.
The constraint
A useful conversation could still end without action.
The decision
Evaluations and pathways saved the work as decisions, milestones, and tasks the founder could return to.
Product Experience
Personalized Pathways
A founder's goal becomes a saved sequence of milestones and tasks they can revisit.
Business Concept Evaluation
The report breaks an idea into questions about demand, feasibility, differentiation, and the business model.
Startup Mentor
The conversation draws on the founder's saved business context and research material.
The outcome
What changed
The result
- ✓Launched AI Startup Coach as a native iPhone app.
- ✓Shipped concept evaluations and personalized pathways with saved tasks and milestones.
- ✓Built the supporting services for AI providers, user data, subscriptions, notifications, and access control.
Looking back
The most useful output was often a small next step. The product became clearer when I designed around what a founder could do after closing the app.
Saved context gave the coaching continuity. It also made the quality of the founder's initial inputs worth designing carefully.
What I'd explore next
I would make sources easier to inspect and update pathways as the founder gathers evidence and completes milestones.
Where I contributed
Product
- Founded the business and took the coaching concept through App Store launch.
Design
- Designed onboarding, concept reports, and pathways around a founder's next decision.
Engineering
- Built the SwiftUI app and integrated AI, persistence, subscriptions, and notifications.
Related projects
Ollie AAC
Founder/CEO
Communication boards that families can create and change for everyday life, built for iPhone and iPad.
SwiftLM
ML Engineer
A Swift coding model I fine-tuned on one GPU, published on Hugging Face, and set up to run locally.
Mozaic App
Senior Product Manager
Making shared earnings easier to understand, from the split agreement to the payout.