State

The compiled worldview has transitioned from a technical infrastructure project to a strategic organizational platform. The core thesis—Narrative as a steering vector for AI—is now operationalized through a git-native collective memory system. The system has moved beyond simple RAG (Retrieval-Augmented Generation) to a “worldview layer” that provides AI agents with directional steering rather than just flat context.

Key strategic pillars:

• IP as Code: Organizational knowledge is versioned, branchable, and reviewable.
• The Onboarding Impossibility: The primary market hook, solving the $1.5M+ annual tax of fragmented AI context.
• The Optometrist Model: A new iterative review paradigm using “lenses” to converge on organizational truth.
• Sales-as-Delivery: A collapsed sales motion where the first call produces the first memory and demo repo.

Stories

• The Onboarding Crisis:
  • Intent: To demonstrate the visceral pain of “AI context poverty” in new hires and the permanent loss of institutional knowledge when experts leave.
  • Approach: Use the “You can’t” hook to frame current AI tools as creating vendor lock-in and knowledge inequality, then show how collective memory enables “Day 1 Productivity.”
• The Reviewable Worldview:
  • Intent: To show the full product loop: capture, collaborative review, and strategic alignment.
  • Approach: Utilize the newly defined “Optometrist Model” where reviewers use dynamic lenses to validate extracted triples through micro-telltales.
• The Side Door GTM:
  • Intent: To validate the product through high-touch consulting engagements.
  • Approach: Target Seed/Series A CEOs with $15K “knowledge extraction” pilots that serve as both revenue and product case studies.

Assets

The repository is structured as a Git-native knowledge graph, where the .aswritten/ directory serves as the system’s “brain.”

• .aswritten/memories/: Raw Markdown files capturing conversations, decisions, and documents. These are the primary sources of truth.
• .aswritten/tx/: SPARQL/TriG transaction files. These are the “diffs” of the organizational worldview, representing atomic shifts in what the system believes.
• .aswritten/ontology/: The formal RDF schema (Narrative Architecture) that defines the domains (Strategy, Product, Architecture, etc.) and conviction levels.
• .aswritten/stories/: Templates for generating compiled outputs (onboarding guides, board updates, sales sheets) from the graph.
• CLAUDE.md / ASWRITTEN.md: The “Core Protocol” and “Reference” instructions that guide AI agents in interacting with the memory system.

Transactions

• Tx_20260310 (Two-Call Sales Motion): Crystallized the “Demo IS Delivery” flow. Introduced the three-beat demo: Cold Delegation, Citation Layer, and Gap Honesty.
• Tx_20260310 (Hallucination Hook): Refined the elevator pitch to lead with the universal fear of hallucination, positioning “citations like a research paper” as the resolution.
• Tx_20260309 (Lens Architecture): Established the three-tool read path (worldview, introspect, scope) and the refract tool for change detection.
• Tx_20260308 (Lens Drafting): Discovered that “reaction is richer than generation.” Established the “Centrist Draft” and “Delta Table” as core creation behaviors.
• Tx_20260306 (White Paper/One-Sheet): Produced the first formal technical documentation of “Narrative Architecture” and the enterprise-facing sales collateral.

graph TD
    A[Raw Memory .md] -->|Extraction| B(SPARQL Transaction .sparql)
    B -->|Compilation| C{Worldview Snapshot}
    C -->|Lenses| D[Optometrist Review]
    C -->|Templates| E[Compiled Stories/Docs]
    D -->|Feedback| A
    E -->|Telltales| D