THIOS: AI AS TECHNICAL CO-FOUNDER
Validating AI-augmented workflows by building a complete product ecosystem solo in 18 months
18mo
Idea to Revenue
70%
Faster Prototyping
5
Shelter Variants
3
Websites Built
The Strategic Hypothesis
How might a senior product leader leverage AI tools to operate at the velocity of a full team? I wanted to test whether an experienced generalist could compress traditional product development stages—strategy, research, design, engineering—into a solo workflow with AI as a scalable collaborator.
The goal wasn't just to build a product. It was to validate a model that any startup could use to run leaner.
Why This Was Hard
This wasn't a software side project—it was a physical product with real constraints:
- Solo execution: No team, no funding—every role from strategy to CAD to code to marketing
- Physical + digital: Not just software—actual shelter systems requiring manufacturing knowledge
- Time compression: 18 months to prove the model from zero to revenue
- Tool evolution: AI capabilities rapidly shifting—had to continuously adapt workflows
- Quality bar: Had to be indistinguishable from team-built work to validate the model
If this only worked because standards were lowered, it wouldn't prove anything useful.
Strategy
I treated AI as a technical co-founder, not a tool. This meant:
- Role-Based Collaboration: Different AI tools for different phases—research, design, development, CAD
- Conversational Development: Iterating through dialogue, not just prompts—building context over time
- Quality-First: Never ship something I wouldn't ship with a team—AI accelerates, doesn't excuse
- Continuous Adaptation: Workflow evolved as tools improved—Copilot → Cursor → Claude Code
Execution
Research & Strategy
Obsidian MD + AI for project management and research synthesis. Built a knowledge base that AI could reference across sessions.
Brand & Visual Design
Figma for visual design with AI-assisted design-to-code workflows. Brand identity inspired by early 2000s open-source culture.
Web Development
Progressed through GitHub Copilot → Cursor → Claude Code as tools matured. Each evolution compressed cycles further. Built 3 production websites.
Physical Product
OnShape for CAD modeling with AI helping translate concepts into precise specifications. 5 shelter variants designed and one physical prototype built.
Results
- Delivered a complete product ecosystem in 18 months: brand, 3 websites, 5 shelter variants, CAD models, physical prototype, revenue-generating handbook
- Compressed prototyping cycles 70% faster through conversational development
- Built a 3D configurator (Three.js + React) that lets customers design custom shelters
- Validated transferable patterns for AI-augmented solo work applicable to product organizations
What This Proves
This work demonstrates:
- That AI-augmented workflows are ready for production use—not just prototypes
- That experienced generalists can operate across boundaries that traditionally required specialists
- That the patterns are transferable—any startup or product team can run leaner with the right approach
The real outcome was proving that the future of product development isn't larger teams—it's better leverage.
Gallery
LET'S TALK
Interested in discussing AI-augmented workflows or exploring opportunities?