2025-11-15
Prototyping as a Tool for Understanding Markets
When you don’t know what a market wants, the best move is not analysis — it’s prototyping. Research feels rigorous, but building exposes constraints that slides and models hide.
Over the past two years I built three distinct prototypes across different domains:
- AI assurance (Assurenest AI)
- GameFi/Web3 (HuskyChain)
- Quant analytics (Factor Lens)
These weren't products in search of exits; they were experiments in search of truth.
What Prototyping Reveals
Prototypes reveal four things faster than any interview, deck, or report:
-
User constraints
Users don’t behave the way founders imagine.
Assurenest AI revealed that AI adoption inside SMEs is gated by compliance and confidence, not just productivity. -
Technical constraints
Cloud reality forced decisions: SSE vs WebSockets, Firebase vs Cloud Run, cold starts vs cost, wallet-last onboarding for HuskyChain, etc. -
Economic constraints
Tokenomics in Web3 and factor models in quant trading both exposed the same rule: incentives determine behavior. -
Compliance constraints
AI assurance lives next to regulation; GameFi lives next to financial incentives; quant lives next to risk.
Compliance shapes products more than features do.
Ideas don’t surface these constraints — implementation does.
Why Prototypes > Decks
A deck tries to persuade.
A prototype tries to learn.
Decks optimize for narrative.
Prototypes optimize for reality.
A prototype forces contact with:
- actual UX
- actual data
- actual infrastructure
- actual economics
- actual users (or players)
- actual regulation or compliance thinking
Founders underestimate how much discovery happens through implementation.
The Hidden Benefits
1. Prototypes Expose Incentives
HuskyChain made this visible. Token design was not financial — it was incentive design. Tiers, NFTs, referrals, and staking all changed user behavior more than “features”.
2. Prototypes Compress Research
Factor Lens turned crypto factor modeling from a theoretical exercise into dashboards and simulations. Once results were visual, research accelerated.
Visualization is a research tool, not a finishing step.
3. Prototypes Surface Compliance Early
Assurenest AI forced questions normally asked at Series A or pre-audit:
- What counts as hallucination?
- How do you log bias?
- Who reviews outputs?
- What is risk?
- Who is accountable?
Compliance is not a feature — it is a boundary condition.
Prototypes Are Filters
Each prototype answered key questions:
- Assurenest AI: Is AI assurance a technical problem or a compliance problem? (Answer: both.)
- HuskyChain: Can Web3 onboard non-crypto users without wallets? (Answer: yes, Telegram proves it.)
- Factor Lens: Are factor models in crypto extractable or just academic? (Answer: partial signal, needs regimes.)
These answers came from building, not theorizing.
Outcomes That Still Count as Wins
Prototypes do not need to become products.
They can produce:
- pivots
- insights
- frameworks
- clarity
- constraints
- proof of skill
- proof of execution
- technical conviction
- investor signal
- visa evidence
- new hypotheses
A “failed” prototype can be intellectually profitable.
Closing Thoughts
Prototyping is not the path to a product; it is the path to understanding. Whether the outcome is a product, a pivot, or shelving the idea, the prototype still pays for itself in insight.
The most valuable part of the past two years wasn’t the code — it was the learning: how AI behaves under audit, how incentives change player behavior, and how data becomes research when it becomes visible.
Prototypes are where markets stop being theory and start being constraints.