AI PM Exploration Toolkit 🎓

The AI Skills Learning Platform for Product Managers
Close your AI PM skills gap through hands-on exploration and practical application. Transform from AI-curious to AI-confident.

Philosophy

Primary Mission: Bridge the AI skills gap that's leaving product managers behind.

As AI transforms product development, strategic PMs need hands-on fluency with AI tools, not just conceptual understanding. This toolkit provides a safe learning environment where you can experiment, fail fast, and build confidence with AI-powered product management techniques.

The Learning Journey:

• 🎓 Education: Personal AI classroom for skill building
• 🔍 Exploration: Discover what's possible through guided experimentation
• 🧪 Experimentation: Learn by doing with real product management scenarios
• 📊 Explanation: Apply new skills to create compelling product narratives

Secondary Benefit: Once you've learned the tools, use them for Proof-of-Life (PoL) Probes — lightweight, disposable tests that de-risk product decisions without burning engineering resources.

Learn First. Apply Second. Lead Confidently.

Get Started in 2 Steps

Ready to turn product ideas into evidence? Let's get you set up.

Step 1: Download the Toolkit

git clone https://github.com/deanpeters/ai-pm-exploration-toolkit.git
cd ai-pm-exploration-toolkit

Step 2: Install & Start

python3 core/installer.py

🎉 You're ready! The web dashboard will be available at http://localhost:3000

What You Actually Get

Currently Implemented & Working:

🌐 Web Dashboard (Immediate Visual Access)

• Audio Transcription Tool - Upload MP3/WAV files for AI-powered PM analysis
• AI Chat Assistant - Brainstorm and analyze with local LLMs
• Data Generation - Create synthetic personas and test data
• Market Research - Company lookup and competitive analysis
• Workflow Automation - Access to Docker-based workflow tools

💻 Command Line Tools

• Audio Processing: python3 shared/audio_transcription.py interview.mp3 --use-case user_interviews
• AI Strategic Chat: python3 shared/ai_chat.py --mode pm_assistant --interactive
• Data Generation: python3 shared/data_generator.py --personas 50 --industry saas
• Market Research: python3 shared/market_research.py --company "CompanyName"

🚀 AI Consultation Methods (Multiple Options)

External AI Systems (Drag & Drop Context)

Perfect for strategic consultation using Claude, ChatGPT, or Gemini:

1. Upload these 6 context files to your AI project:
   • AI_CONSULTATION_CONTEXT.md
   • TECHNICAL_SPECIFICATION.md
   • CONSULTATION_TEMPLATES.md
   • TOOLKIT_CONFIG_REFERENCE.json
   • SAMPLE_OUTPUTS.md
   • QUICK_CONTEXT_SETUP.md
2. Use ready-made templates for common PM scenarios
3. Get strategic guidance with full toolkit context

Native Goose CLI Integration (Local AI Agent)

Autonomous AI analysis with direct toolkit access:

# Start strategic analysis session
goose session --name pm_strategy_analysis

# Pre-built workflows for common PM challenges
goose session --name feasibility_check
goose session --name competitor_analysis
goose session --name feature_prioritization

VS Code + Continue Integration (Developer-Friendly)

AI assistance directly in your development environment:

# Open toolkit in VS Code
code /Users/deanpeters/ai-pm-exploration-toolkit

# Use Continue shortcuts:
# Ctrl+I: Ask about implementation
# Ctrl+L: Strategic AI discussions
# Ctrl+K: Generate PM documentation

Visual Documentation (All Users)

Beautiful markdown reading experience with MarkText:

# Enhanced documentation viewing (requires MarkText installation)
open -a MarkText AI_CONSULTATION_CONTEXT.md

# Or use the aipm_marktext alias if setup.sh was run
aipm_marktext AI_CONSULTATION_CONTEXT.md

• python3 shared/audio_transcription.py - Speech-to-text with PM insights
• python3 shared/ai_chat.py - Interactive AI conversations
• python3 shared/market_research.py - Company research automation
• python3 shared/data_generator.py - Synthetic data creation
• ./orchestrate-workflows.sh - Docker workflow management

🎙️ Audio Intelligence System (Phase 7.1 Complete)

• OpenAI Whisper Integration - Local speech-to-text processing
• 6 PM Workflow Templates:
  • User Interview Analysis
  • Stakeholder Meeting Summaries
  • Product Demo Feedback
  • Competitive Research Processing
  • PM Voice Memo Conversion
  • Customer Support Analysis
• Smart Insight Extraction - Pain points, features, decisions automatically identified
• Multiple Audio Formats - MP3, WAV, M4A, FLAC support

🤖 AI Chat & Analysis

• Local LLM Integration - Works with Ollama (qwen2.5, deepseek-r1, llama3.2)
• PM-Specific Modes - Strategic analysis, brainstorming, competitive analysis
• Conversation Management - Save and resume discussions
• Model Auto-Detection - Automatically finds available AI models

📊 Data & Research Tools

• Synthetic Data Generation - Create realistic user personas and datasets
• Market Research Engine - Company information lookup and analysis
• Workflow Automation - Docker-based n8n and Langflow (ToolJet, Typebot, Penpot have compatibility issues)

🦢 Goose CLI Integration (Phase 7 Complete)

• Alternative AI Assistant - Claude Code-like experience at lower cost
• Multi-step Task Automation - Complex PM workflows with AI guidance
• Local Configuration - Works with your existing Ollama setup

Core Tools Status

Tool	Status	Web Access	CLI Access	Description
Audio Intelligence	✅ Production	✅ Upload Interface	✅ Full CLI	Whisper-powered transcription with PM analysis
AI Chat System	✅ Production	✅ Chat Interface	✅ Full CLI	Local LLM for PM brainstorming
Data Generation	✅ Production	✅ Web Form	✅ Full CLI	Synthetic personas and datasets
Market Research	✅ Working	✅ Company Lookup	✅ Full CLI	Basic company research
Workflow Automation	⚠️ Partial	✅ Status Page	✅ Docker Scripts	n8n and Langflow (others broken)

Real Example Commands

Audio Processing:

# Process a user interview
python3 src/audio_transcription.py interview.mp3 --use-case user_interviews

# Get PM workflow options
python3 src/pm_audio_workflows.py --list

AI Chat:

# Start interactive PM assistant
python3 src/ai_chat.py --mode pm_assistant --interactive

# Analyze with specific model
python3 src/ai_chat.py --mode analysis --model qwen2.5

Data Generation:

# Generate user personas
python3 src/data_generator.py --personas 50 --industry saas

# Create survey responses
python3 src/data_generator.py --surveys 100 --topic "product satisfaction"

Research & Workflows:

# Look up company information
python3 src/market_research.py --company "Notion"

# Start workflow tools
workflows/orchestrate-workflows.sh start

Architecture & Files

Clean Directory Structure:

ai-pm-exploration-toolkit/
├── 📁 core/                          # System essentials
│   ├── installer.py                   # Main installer
│   ├── toolkit.json                   # Tool configuration
│   └── run_tests.sh                   # System validation
├── 📁 src/                          # Core implementations
│   ├── audio_transcription.py         # Audio processing engine (678 lines)
│   ├── pm_audio_workflows.py          # PM workflow templates (775 lines) 
│   ├── ai_chat.py                     # AI chat system (613 lines)
│   ├── market_research.py             # Research tools (400+ lines)
│   └── data_generator.py              # Data generation (300+ lines)
├── 📁 web/                          # Web dashboard
│   ├── app.py                         # Flask server
│   ├── templates/                     # HTML interfaces
│   └── tools/                         # Web tool pages
├── 📁 workflows/                    # Docker orchestration
│   └── orchestrate-workflows.sh       # Container management
├── 📁 outputs/                      # All generated content (organized!)
│   ├── research/                     # Company research & analysis
│   ├── transcripts/                  # Audio processing results
│   ├── personas/                     # Generated user data
│   ├── conversations/                # AI chat logs
│   └── reports/                      # Analysis reports
├── 📁 docs/                         # User documentation
├── README.md                         # This file
└── CLAUDE.md                         # Project context

Installation Requirements

• Python 3.8+ with pip
• 10 GB free disk space
• Internet (for initial setup)
• Docker (optional, for workflow tools)
• Git (for installation)

Platform Support:

• ✅ macOS (Intel & Apple Silicon)
• ✅ Linux (Ubuntu 20.04+)
• ⚠️ Windows (basic support)

Quick Test

After installation, verify everything works:

# Quick system validation
core/run_tests.sh --quick

# Test core systems individually
python3 src/audio_transcription.py --status
python3 src/ai_chat.py --status  
python3 web/app.py &  # Starts web server

# Open browser to http://localhost:3000

What's Next

Phase 8 Planned Features:

• More workflow tool integrations
• Advanced prompt engineering tools
• Real-time collaboration features
• Extended AI model support

The 4E Framework in Practice

🎓 Education: Learn Through Doing

• Use the web dashboard to upload audio and see AI analysis
• Experiment with different AI models through the chat interface
• Generate synthetic data to understand AI capabilities

🔍 Exploration: Discover Possibilities

• Try different audio workflow templates
• Test AI chat with various PM scenarios
• Explore market research capabilities

🧪 Experimentation: Test with Data

• Process real user interview recordings
• Generate personas for your product scenarios
• Use AI to analyze competitive intelligence

📊 Explanation: Create Compelling Stories

• Use processed insights in stakeholder presentations
• Generate supporting data for product decisions
• Create evidence-based narratives with AI assistance

Contributing

We welcome contributions that enhance the PM learning and exploration experience:

• Tool improvements - Enhance existing functionality
• New integrations - Add tools that fit the 4E framework
• Documentation - Help other PMs succeed
• Bug fixes - Make the toolkit more reliable

See CONTRIBUTORS_TEST_PLAN.md for testing requirements.

Philosophy in Action

Traditional PM approach:

1. Write requirements
2. Wait for development
3. Launch and hope
4. Learn after release

AI PM Exploration approach:

1. Generate synthetic data
2. Process user insights with AI
3. Create evidence-based narratives
4. De-risk decisions before committing resources

---

🎯 Show before Tell. Touch before Sell.
💡 Use the cheapest prototype that tells the harshest truth.

Built for Product Managers who need evidence, not opinions.