AI-powered assumption validation tool that helps founders identify hidden risks in 30 seconds.

Try it live →

80% of startups fail not because they built the wrong thing, but because they never tested the right assumptions. Founders spend months building products based on untested beliefs, only to discover critical flaws too late.

Common blind spots:

• "Users will pay $15/month" ← Did you test willingness to pay?
• "We can acquire customers for <$50" ← What's your actual CAC?
• "The technology works at scale" ← Have you stress-tested it?

Blindspot Spotter uses AI to:

1. Break down your idea into 6-8 testable assumptions using first principles
2. Score each assumption on RISK (impact if wrong) and TESTABILITY (cost/time to validate)
3. Map assumptions visually on a 2×2 matrix showing which to test first
4. Generate practical experiments ($0-$3K, 1-4 weeks) with success criteria

• AI Analysis: Describe your idea, get AI-generated assumptions
• Manual Entry: Input your own assumptions for analysis

Visual matrix with four quadrants:

• Test Now (high risk + easy to test) → Top priority
• Critical Risk (high risk + hard to test) → Dangerous blind spots
• Quick Wins (low risk + easy to test) → Nice-to-haves
• Defer (low risk + hard to test) → Lowest priority

Each assumption includes:

• 3-sentence test method
• Realistic cost estimate ($0-$10K)
• Time to complete (3 days - 12 weeks)
• Success criteria with concrete metrics

AI flags 2-3 dangerous assumptions founders typically miss

Validate your startup idea before building. Use when:

• Exploring a new product concept (0→1)
• Pivoting to a new direction
• Preparing for investor conversations
• Planning your MVP roadmap

De-risk new feature designs. Use when:

• Designing greenfield features from scratch
• Creating first-time user flows (onboarding, etc.)
• Entering new user segments
• Planning UX research sprints

Note: Best for NEW ideas where you have no data yet. If you have analytics or user research, use that instead.

Input (AI Mode):

"AI-powered scholarship matching platform for high school students"

Output:

• 7 assumptions identified
• 3 marked as "Critical Risk" (high impact + expensive to test)
• 4 marked as "Test Now" (high impact + cheap to validate)
• Practical experiments like: "Interview 15 high school counselors to validate referral model assumption - Cost: $0, Time: 2 weeks"

• Framework: Next.js 14 + TypeScript
• AI: Anthropic Claude Sonnet 4 API
• Styling: Tailwind CSS + shadcn/ui
• Deployment: Vercel (auto-deploy from GitHub)
• Charts: Custom SVG-based 2×2 matrix visualization

1. First Principles Decomposition Breaks product ideas into core assumptions across:

• User Behavior (will people use/pay?)
• Market Dynamics (does the market work this way?)
• Technical Feasibility (can we build it?)
• Business Model (do the economics work?)
• Operations (can we deliver sustainably?)

2. Risk Scoring (1-10)

• 9-10: Critical - entire business fails if wrong
• 7-8: High - significantly impacts growth/economics
• 5-6: Moderate - affects efficiency but adaptable
• 1-4: Low - nice-to-have optimizations

3. Testability Scoring (1-10)

• 8-10: Easy - test in 1-2 weeks with <$1000
• 5-7: Moderate - takes 2-4 weeks, $1000-$3000
• 1-4: Hard - requires 4-12 weeks, $3000-$10,000

4. Experiment Design Provides concrete, low-cost validation methods:

• Landing pages + ads
• User interviews
• Concierge MVPs (manual service)
• Small pilots with 20-50 users
• Competitive analysis

• Node.js 18+
• npm or yarn
• Anthropic API key

# Clone the repository
git clone https://github.com/sonakshiv10-aws/blindspot-spotter.git

# Navigate to project
cd blindspot-spotter

# Install dependencies
npm install

# Set up environment variables
cp .env.example .env.local
# Add your Anthropic API key to .env.local

# Run development server
npm run dev

Open http://localhost:3000 to see the app.

# Run development server
npm run dev

# Build for production
npm run build

# Start production server
npm start

# Run linter
npm run lint

blindspot-spotter/
├── src/
│   ├── pages/
│   │   ├── index.tsx          # Main landing page with input modes
│   │   └── api/
│   │       └── analyze.ts     # API route for Claude integration
│   ├── components/
│   │   └── AnalysisMatrix.tsx # 2×2 matrix visualization
│   └── styles/
├── public/
└── package.json

v1 (Current):

• ✅ AI-powered assumption generation
• ✅ Manual assumption entry
• ✅ Risk/testability scoring
• ✅ Visual 2×2 matrix
• ✅ Experiment recommendations

v2 (Future):

• Save/export analysis as PDF
• Share via unique URL
• Track tested vs. untested assumptions
• Integration with experiment tracking tools
• Team collaboration features

This is a personal portfolio project, but feedback is welcome!

Found a bug? Open an issue on GitHub Have a suggestion? DM me on LinkedIn

MIT License - feel free to fork and modify for your own use.

Built by Sonakshi Verma as part of a career transition from architecture → product management.

Why I built this: After spending 6 months building a product based on untested assumptions, I learned the hard way that validation beats vision. This tool automates the assumption mapping process I wish I'd had from day one.

Tech used: Next.js, TypeScript, Claude AI, Tailwind CSS Built in: 2 weeks (November 2025)

• LinkedIn: sonakshiverma
• Email: sonakshiv10@gmail.com
• Live Demo: blindspot-spotter.vercel.app

⭐ If this tool helped you avoid a costly mistake, consider starring the repo!