Strategy OS

An AI-Augmented "Playing to Win" Strategy System

This repository contains a living, breathing business strategy operating system. It combines the rigorous "Playing to Win" framework with modern AI agent workflows to create a strategy that is dynamic, traceable, and easily updatable.

Core Concepts

1. The Framework: "Playing to Win"

The strategy is structured around the five integrated choices of the "Playing to Win" cascade (Lafley & Martin):

1. Winning Aspiration: What is our winning purpose?
2. Where to Play: Where will we compete?
3. How to Win: How will we win where we have chosen to play?
4. Capabilities: What capabilities must be in place?
5. Management Systems: What systems are required?

2. The Agentic Workflow

This isn't just a folder of documents. It is a workspace designed for collaboration between a Human Strategist and an AI Agent.

• Source of Truth: The files in /strategy are the authoritative record.
• Inbox-Driven Updates: New information (articles, data, ideas) is dropped into /inbox and processed systematically.
• Traceability: Every change to the strategy is linked to source materials and documented in /archive.

Directory Structure

Directory	Purpose
/strategy	The Source of Truth. Contains the 5 core Markdown files defining the current strategy.
/inbox	Drop raw files here (PDFs, notes, articles). The Agent monitors this folder to initiate processing.
/wiki	A React-based static site generator that turns the /strategy folder into a beautiful, navigable website.
/skills	Knowledge base for the Agent. Contains methodologies, templates, and "how-to" guides for strategic tasks.
/archive	Historical record. Processed inbox items and superseded strategy files move here.
/outputs	Generated deliverables (presentations, reports) derived from the core strategy.

How It Works

The system operates on three primary workflows:

1. Inbox-Driven (Human-First)

For processing existing materials.

1. Drop a file in /inbox (reports, notes, data).
2. Agent Analysis: The Agent analyzes the file and proposes an Implementation Plan.
3. Update: Upon approval, the Agent updates the /strategy files and rebuilds the wiki.

2. Autonomous (AI-First)

For filling logical gaps.

1. Identify Gap: The Agent identifies missing data or logical inconsistencies.
2. Research: The Agent conducts web research or internal analysis.
3. Propose: Findings are presented as a research artifact for human review.

3. The Interview Protocol (Collaborative)

For extracting strategic wisdom.

1. Formulate: The Agent researches context and formulates 1-5 "Beautiful Questions."
2. Interview: The Agent interviews the Human Strategist to extract "dark data" and judgment.
3. Synthesize: The dialogue is converted into formal strategy updates.

Agent Instructions (AGENT.md)

The file AGENT.md in the root directory contains the "operating system" for the AI Agent. It defines:

• Boot sequence and priorities.
• Protocols for handling IP (Client vs. Firm).
• Detailed workflows for processing materials and resolving conflicts.

---

Built for the future of strategic consulting.