Define the product before you build the company.
Deep tech and AI-native founders rarely lose on technology. They lose on definition — an inability to show where they sit in the value chain, what they must be best at, and how the system fails under load.
First-Principles Product is the structured method we use to fix that: a sequence of design tools that turn a raw idea into a defensible, investable product definition. Operator-tested, modernized for AI-native builders.
Mis-built products. Teams build before they define, then discover the real customer was never on the map.
No articulated moat. Strong technology, weak story — the value-chain position is never made legible to capital.
Structured clarity. A traceable line from customer need to architecture to failure mode — defensible under diligence.
The methods library.
Six structured design tools, each explained in two minutes. They run in sequence — front of the workflow to the end — so the library itself teaches the design flow. Use one, or use them in order.
Map your value chain before you build.
Identify every stakeholder in your ecosystem — not just the end user — and what flows between them: money, data, information, influence. Who you leave off the map is a value judgment.
Decide what to build first.
Align the team on what truly matters before a line of code is written. Rank competing efforts on impact and feasibility, and kill the work that fails both.
Turn what customers want into what you must measure.
The House of Quality translates fuzzy customer needs into specific, measurable requirements — and makes every trade-off visible instead of buried.
Find where cost and value diverge.
Decompose the product into the functions it must perform, then expose where you are spending money on things the customer does not actually value.
Choose between architectures without the politics.
Generate solution variants systematically, then score them against a reference baseline. A rigorous way to pick an architecture, a vendor, or a build-vs-buy path.
Know how it breaks before it does.
Systematically surface what can go wrong, how likely, how severe, how detectable — and prioritize fixes by risk. Reactive becomes proactive.
Forged at Stanford. Rebuilt for the AI Era.
These methods are decades-proven in hardware and manufacturing. The discipline holds; the surface area changed. Each tool now carries a modern angle for software, services, and AI systems.
- CVCA
Beyond the algorithm. Model providers, data labelers, and regulators (EU AI Act, FTC) belong on the map.
- QFD
Evaluation metrics. Map user needs to the evals and acceptance criteria that actually prove the system works.
- PUGH
Concept selection. Feature or a company? Build vs. buy vs. partner — scored, not argued.
- FMEA
Failure modes. From scenario planning and ethical nightmares to a responsible-AI principle stack and governance.
Bring the method to your team.
Delivered to accelerators, founder cohorts, and corporate product teams across the Silicon Valley ↔ Tokyo ↔ Taipei corridor. Two formats.
The single-method talk.
A focused session on any one method in the library — the highest-leverage way to introduce structured design to a cohort or a team that has never seen it.
- One method, worked through a live example
- A hands-on artifact participants build in the room
- Fits a demo-day prep slot or a lunch-and-learn
The Structure-to-Scale intensive.
The complete structured-definition sprint: a cohort takes a real product from value-chain map to risk analysis, leaving with a defensible definition they can take to capital.
- All core methods, run in design-flow order
- Teams apply each tool to their own venture
- Capstone: a design story from customer need to launch call
The toolkit is drawn from Stanford's structured engineering-design tradition — among the most rigorous methods taught in graduate product design — and carried into practice through operator work in semiconductors, hardware, and AI. We have modernized it for AI-native builders, but the discipline is the same one that has produced defensible products for forty years. First-principles, not first-draft.
Define it once. Defend it everywhere.
Bring First-Principles Product to your venture or your cohort — as a strategy session, a single talk, or the full intensive.