Skip to main content

Truth Management

Truth management is short for operational truth management: the discipline of making your operational reality agent-legible. It's the systematic work of documenting and organizing the conceptual frameworks that guide decision-making for individuals and organizations, so that shared vocabularies, explicit principles, and coherent worldviews enable aligned action by both humans and AI agents.

A note on terminology

We say "truth management" for brevity, but the full term is "operational truth management," aka agent-legible operational reality. We are not claiming to manage capital-T metaphysical Truth. We're managing the operational reality an entity runs on (its working assumptions, decision criteria, and the premises it uses to navigate the world) and putting it into a form that both humans and AI agents can read, query, and act on.

Why Call It "Truth" and Not "Context"?​

This discipline could just as easily be called "context management," and in a technical sense that's exactly what it is: curating the context lake that agents and team members draw from.

We use "truth" deliberately to convey the gravity of the work. Every file in that lake becomes a premise your agents and teammates will act on. Sloppy context produces sloppy action at machine speed. Treating it as "truth" (something that has to be accurate, current, and coherent) is the posture the work requires.

Why This Matters​

To change the world meaningfully, individuals and organizations must be deeply grounded in a foundation of truth: their assumptions about reality, motivating narrative, and decision-making frameworks. Without this grounding, individuals operate from unconscious contradictions and organizations suffer from misaligned assumptions that create coordination failures. The accumulated version of that failure is agentic OS debt.

Truth management solves this by systematically documenting and evolving shared understanding. When truth is made explicit through version-controlled documentation, scattered assumptions are replaced by coherent foundations for aligned action for every human and AI involved. This is the substrate agentic strategy runs on and the precondition for a self-improving enterprise.

Read the full argument for the three interconnected claims that build toward this thesis.

Ready to start? See the getting started guide for a sequenced Day 1 through Week 4 plan.

The Framework​

Principles​

Core operating principles for effective truth management, ordered from foundational claim to structural commitments:

  • Don't Assume Common Sense: The foundational claim. Remove "common sense" from your vocabulary; only shared understanding exists.
  • Align Before Committing: The partnership-level application. Never start meaningful projects without explicit, documented alignment.
  • Make Every File Count: The per-file quality bar. Every document must actively support right action with minimal noise.
  • Make Your Company Refactorable: The structural requirement. Can you grep and edit your entire company OS with an agent call? If not, why?
  • Empower Your Truth Manager: The organizational commitment. Truth management requires real authority to enforce coherence.
  • Protect Your Truth: The security layer. Match access controls to sensitivity levels.

Processes​

Systematic workflows for implementing truth management:

  • Start Your Company Bible - Build comprehensive documentation capturing your organization's "way"
  • Migrate to Refactorable Systems - Move from siloed tools to grep-able, version-controlled formats
  • Truth as Context - Ensure AI agents have full organizational context when creating or modifying documentation
  • Maintain Coherence - Documents decay by default. Coherence is an active practice: automated checks, cross-linking, and scheduled audits that keep your truth repository reflecting current reality

Tools​

  • Source Controller - Version control platforms (e.g., GitHub) that store and evolve your truth repository
  • Voice Transcriber - Speech-to-text tools that lower the friction between thinking and documenting

Truth Management and Field Notes​

Truth management is the discipline. Field notes are what it produces.

Without truth management, living documentation degrades into a wiki: well-intentioned at first, then gradually filled with outdated, contradictory, and unreliable information. Truth management imposes the rigor that prevents this: version control, explicit ownership, systematic review, and the principle that every file must actively support right action.

We live in an age of embellishment where social media algorithms reward hype over accuracy, and the traditional textbook industry is structurally misaligned with the interests of learners. The Applied AI Society's documentation exists as a safe, reliable home for the truth: field notes from practitioners doing the work, managed with the discipline described on this page, so that when we educate people, we educate them from love, not propaganda.

About​

This framework was created by Gary Sheng, founder of the Applied AI Society. It emerges from observing that competitive advantage increasingly comes from organizational capability rather than just proprietary technology. As AI agents become integral to how work gets done, organizations with well-documented truth gain compounding advantages: their agents operate from the same foundations as their humans, enabling coordinated action at unprecedented speed.

Truth management as a discipline is still nascent territory. The Applied AI Society is committed to doing this right from day zero: building our entire organizational infrastructure on these principles from the start, rather than retrofitting them later. Our workspace is a living implementation of this framework, and we're learning and evolving these ideas in real time.

  • Why Field Notes: What a truth management discipline produces for a community
  • Canon: The tenets that truth management operationalizes
  • North Star: The direction every file in a truth repository should advance