Shape Up AI Native

Complete guide to adapting Basecamp's Shape Up methodology for AI-native software development teams

⚠️ Important Notice

This is NOT an official Basecamp project.

This is an adaptation of the original Shape Up methodology by Basecamp, reimagined for the age of AI coding agents.

By Sérgio Lindolfo Ferreira — 30+ years building software, now exploring how Shape Up principles evolve when AI agents join your development team.

🎯 What is Shape Up AI Native?

Shape Up (2019) revolutionized product development with:

• 6-week cycles instead of endless sprints
• Shaping work before building
• Betting instead of backlogs
• Team autonomy and accountability

But in 2026, the game has changed:

• AI agents build code 10-100x faster than humans
• Developers review instead of write
• Cycles compress from weeks to days
• The bottleneck shifts from coding to validation

Shape Up AI Native adapts these timeless principles for AI-augmented teams.

🔄 Workflow at a Glance

%%{init: {'theme':'base', 'themeVariables': { 'primaryColor':'#FFE5B4','secondaryColor':'#E3F2FD','tertiaryColor':'#F3E5F5','primaryBorderColor':'#FF6B6B','secondaryBorderColor':'#4ECDC4','tertiaryBorderColor':'#95E1D3'}}}%%

graph TD
    Start([🎯 Shape Up AI Native Workflow]):::startNode
    
    %% SHAPING PHASE
    Start --> Shaping[🔨 SHAPING<br/>Basecamp]:::shapingNode
    Shaping --> ShapingWork[Problem Definition<br/>+ Appetite<br/>+ Solution Sketch]:::workNode
    ShapingWork --> PitchOut[📄 pitch.md]:::artifactNode
    
    %% BETTING PHASE
    PitchOut --> Betting[🎲 BETTING<br/>Basecamp]:::bettingNode
    Betting --> BettingWork[Review Pitches<br/>+ Commit Resources<br/>+ Assign Teams]:::workNode
    BettingWork --> ApprovedPitch[✅ Approved Pitch]:::artifactNode
    
    %% BUILDING PHASE (Claude Code)
    ApprovedPitch --> Building[🚀 BUILDING<br/>GitHub + Claude Code]:::buildingNode
    
    %% SPEC MODE
    Building --> SpecMode[📝 SPEC MODE]:::claudeNode
    SpecMode --> SpecWork[Pitch → Technical Spec<br/>Requirements + Constraints<br/>Architecture Decisions]:::claudeWork
    SpecWork --> SpecOut[📋 spec.md]:::artifactNode
    
    %% PLAN MODE
    SpecOut --> PlanMode[📋 PLAN MODE]:::claudeNode
    PlanMode --> PlanWork[Spec → Task Breakdown<br/>Define Scopes<br/>Estimate Complexity]:::claudeWork
    PlanWork --> PlanOut[📑 plan.md<br/>+ scopes/]:::artifactNode
    
    %% IMPLEMENT MODE
    PlanOut --> ImplMode[💻 IMPLEMENT MODE]:::claudeNode
    ImplMode --> ImplWork[Plan → Code<br/>TDD + Reviews<br/>Pull Requests]:::claudeWork
    ImplWork --> ImplOut[✨ Working Features<br/>+ Tests]:::artifactNode
    
    %% SHIPPING PHASE
    ImplOut --> Shipping[📦 SHIPPING<br/>GitHub Actions]:::shippingNode
    Shipping --> ShipWork[Code Review<br/>+ Merge to Main<br/>+ Auto Deploy]:::workNode
    ShipWork --> Release[🎉 Azure App Services]:::releaseNode
    
    %% Cooldown feedback loop
    Release -.->|Learnings & Bugs| Shaping
    
    %% STYLE DEFINITIONS
    classDef startNode fill:#4A90E2,stroke:#2E5C8A,stroke-width:3px,color:#fff
    classDef shapingNode fill:#FF6B6B,stroke:#C92A2A,stroke-width:3px,color:#fff
    classDef bettingNode fill:#FAB005,stroke:#F08C00,stroke-width:3px,color:#000
    classDef buildingNode fill:#51CF66,stroke:#2F9E44,stroke-width:3px,color:#fff
    classDef claudeNode fill:#845EF7,stroke:#5F3DC4,stroke-width:3px,color:#fff
    classDef shippingNode fill:#FF922B,stroke:#E8590C,stroke-width:3px,color:#fff
    classDef workNode fill:#E3F2FD,stroke:#64B5F6,stroke-width:2px,color:#000
    classDef claudeWork fill:#F3E5F5,stroke:#BA68C8,stroke-width:2px,color:#000
    classDef artifactNode fill:#FFF9DB,stroke:#FFD43B,stroke-width:2px,color:#000
    classDef releaseNode fill:#69DB7C,stroke:#37B24D,stroke-width:3px,color:#fff

Key insight: What Shape Up calls "Building" becomes 3 distinct phases (Spec → Plan → Implement) when working with AI agents. This gives agents the clarity they need at each step.

Want the details? See Development Workflow Guide for the complete step-by-step process.

🚀 Quick Start: Choose Your Path

🆕 New to Shape Up?

Start here: Understand the methodology first

1. Principles — What is Shape Up AI Native?
2. Original Shape Up Book — Read the original (2-3 hours)
3. Come back and see how it adapts for AI agents

🏗️ Ready to Implement?

Follow this path: From zero to shipping with AI agents

1. Create Your AI Programmer — Build an AI agent from scratch (Mac mini + OpenClaw + Claude)
2. Set Up Basecamp — Configure Shape Up workflow in Basecamp
3. Shape Your First Pitch — Learn to shape work for AI agents
4. Run Your First Cycle — Agent builds, you review
5. Real Example — See a complete shaped pitch

📖 Using It Already?

Deep dives and references:

• Shaping Guide — Master the art of shaping for AI
• Betting Guide — Prioritize when agents move fast
• Building Guide — Agent + human collaboration workflow
• Tools Guide — Basecamp, GitHub, CI/CD setup
• Templates — Copy-paste templates for pitches, scopes, kickoffs

📚 Complete Documentation

🎓 Understanding Shape Up AI Native

Core concepts and philosophy:

Total: ~100 minutes to understand the full methodology

🛠️ Implementation Guides

Step-by-step setup instructions:

Total: ~2.5 hours to go from zero to ready

What you'll have:

• ✅ Working AI agent on Mac mini
• ✅ GitHub account + repository access
• ✅ Basecamp project configured
• ✅ Ready to shape and ship

📝 Templates & Examples

Copy these to start fast:

🏢 Who's Using This

ExampleCompany — Software engineering & applied AI company in Portugal

• 4-person technical team + AI agents
• Building a real estate management platform
• Using 1-2 week cycles with continuous PR review

[Your Company Here?] — Share your story

🔑 Key Differences from Original Shape Up

What stays the same:

• ✅ Shaping before building
• ✅ Fixed time, variable scope
• ✅ Betting, not backlogs
• ✅ Circuit breaker (kill projects that don't ship)
• ✅ Appetite-driven development
• ✅ Team autonomy within boundaries

💡 How It Works: The Full Cycle

1. 🔨 Shaping (Continuous)

What: Define problems and rough solutions before committing

Who: Senior devs, product leads, founders

Output: Shaped pitches ready to bet on

Guide: Shaping for AI Agents

Template: Pitch Template

Example: Performance Dashboard Pitch

2. 🎲 Betting (Start of Cycle)

What: Choose what to build next based on review capacity

Who: Decision makers (CTO, founders, senior team)

Time: 1-2 hours at cycle start

Output: Commitments for the cycle

Guide: Betting When Agents Move Fast

Key question: Can we review this many PRs this cycle?

3. 🚀 Building (1-2 Weeks)

What: AI agent implements, human reviews and deploys

Who:

• Agent: Writes code, creates PRs
• Human: Reviews, tests, merges, deploys

Process:

1. Agent reads shaped pitch
2. Creates branch, implements feature
3. Opens PR with description
4. Human reviews (async, non-blocking)
5. Agent responds to feedback
6. Human merges when satisfied
7. Repeat until feature complete

Guide: Agent + Human Workflow

4. 🧹 Cool-Down (1-2 Days)

What: Review, cleanup, prepare next cycle

Activities:

• Review what shipped (vs. what was bet)
• Fix critical bugs
• Refactor small things
• Shape new pitches
• Prepare for betting

Guide: Basecamp Implementation

🎓 Who Should Read This?

👨‍💼 CTOs & Founders

Read: Principles → Betting → Agent Setup Guide

Why: Understand strategic shift from build capacity to review capacity

👨‍💻 Developers

Read: Building → Shaping → Basecamp Implementation

Why: Learn new role: reviewer + deployer instead of writer

📊 Product Managers

Read: Shaping → Principles → Examples

Why: Shaping becomes MORE critical when agents build fast

🏗️ Team Leads

Read: Agent Setup Guide → Basecamp Implementation → Tools

Why: Practical setup and team integration

🆕 Shape Up Newcomers

Read: Original Shape Up Book → Principles → Shaping

Why: Understand foundation before adaptations

🤝 Contributing

We're building this in the open. Your experiences and insights help everyone!

Ways to contribute:

• 📖 Share your experience: How is Shape Up AI Native working for you?
• 📊 Submit case studies: Real examples from your team
• 📝 Improve docs: Found something unclear?
• 🌐 Translate: Help non-English speakers
• 🎨 Add templates: Share what works for you

See CONTRIBUTING.md for guidelines.

📖 Further Reading

Original Shape Up

• Shape Up book — Free, online, ~2 hours
• Introduction — Quick overview
• Basecamp blog — Additional context

AI-Native Development

• Examples — Real pitches and implementations

Community

• GitHub Discussions — Ask questions, share experiences
• Issues — Report problems, suggest improvements

🗺️ Navigation Map

README.md (you are here)
│
├── Understanding
│   ├── docs/principles.md              → What changes, what stays
│   ├── docs/shaping.md                 → Shape work for AI agents
│   ├── docs/betting.md                 → Prioritize when fast
│   ├── docs/building.md                → Agent + human workflow
│   ├── docs/development-workflow.md    → Pitch → Code (Spec/Plan/Implement)
│   └── docs/tools.md                   → Basecamp + GitHub setup
│
├── Implementation
│   ├── docs/agent-setup-guide.md       → Create AI programmer
│   └── docs/basecamp-implementation.md → Set up workflow
│
├── Templates
│   ├── templates/pitch-template.md     → Shape new features
│   ├── templates/spec-template.md      → Pitch → Technical spec
│   ├── templates/plan-template.md      → Spec → Task breakdown
│   ├── templates/pr-template.md        → Pull Request descriptions
│   ├── templates/scope-template.md     → Break work into chunks
│   └── templates/cycle-kickoff-template.md → Start cycles
│
└── Examples
    └── examples/performance-dashboard/ → Complete real example
        └── pitch.md

Quick links:

• 🆕 New? → Principles
• 🏗️ Ready to build? → Agent Setup + Basecamp Setup
• 📖 Already using? → Shaping, Betting, Building
• 💻 Want details? → Development Workflow (Pitch → Code step-by-step)

📜 License

This work is licensed under a Creative Commons Attribution 4.0 International License.

You are free to:

• Share — copy and redistribute in any medium or format
• Adapt — remix, transform, and build upon the material
• Commercial use — for any purpose, even commercially

Under the following terms:

• Attribution — Give appropriate credit to Sérgio Lindolfo Ferreira and link to the license

Repository code examples (if any): MIT License

🌟 About the Author

Sérgio Lindolfo Ferreira

• 30+ years in software development
• Founder of ExampleCompany (Portugal) and ProductCompany (Estonia)
• Early adopter of Shape Up (since 2012)
• Now exploring AI-native development with coding agents

Connect:

• LinkedIn
• GitHub

Based on the original Shape Up methodology by Basecamp.

Built by: Vasco Gama (AI agent) + Sérgio Ferreira (human)

Made with ❤️ in Portugal 🇵🇹