Skip to content

Value Proposition

Documentation at scale requires more than a language model and a prompt. It requires orchestration -- coordinating specialized agents, enforcing quality standards, maintaining governance, and producing output that is consistent, traceable, and auditable. The FCC Agent Team Framework delivers this orchestration through four value pillars.

Pillar 1: Speed

FCC accelerates documentation production by parallelizing work across specialized personas. Instead of a single author (human or AI) writing sequentially, FCC decomposes work into research, creation, and critique phases, with multiple personas operating within each phase.

What this means in practice:

  • A single fcc generate-docs command produces up to 1,348 documentation files from persona specifications and Jinja2 templates.
  • Workflow graphs define the execution order, eliminating manual coordination overhead.
  • The simulation engine can run entire documentation workflows in seconds (mock mode) or minutes (AI mode), compared to days or weeks for manual authoring.
  • Scenario-based execution means the same workflow can be applied to different documentation targets without reconfiguration.

Pillar 2: Quality

Quality in FCC is not aspirational -- it is enforced. Every persona has defined deliverables, and every deliverable must pass through quality gates before it advances through the workflow.

What this means in practice:

  • 25 quality gates define explicit pass/fail criteria for every persona's output. The Research Inventory Completeness gate, for example, requires a capability matrix, annotated references, a traceability matrix, and capability tags -- all present and valid.
  • Quality gate thresholds are configurable. Some gates require 100% compliance (research, governance); others accept 75% for iterative refinement (UI mockups, metrics dashboards).
  • The Critique phase is not optional. The Documentation Evangelist, Blueprint Validator, and Anti-fact Mitigation Specialist are structural parts of the workflow, not afterthoughts.
  • Each persona's R.I.S.C.E.A.R. specification defines constraints that prevent scope creep, style drift, and output inconsistency.

Pillar 3: Governance

FCC builds governance into the workflow rather than bolting it on after the fact. Three governance personas and a structured compliance layer ensure that AI-generated content meets organizational and regulatory requirements.

What this means in practice:

  • The Data Governance Specialist enforces API contract documentation, data flow compliance, and configuration versioning.
  • The Privacy Taxonomy Engineer classifies data types, maps privacy regulations, and documents handling guidelines.
  • The Anti-fact Mitigation Specialist applies confidence thresholds, verifies source attribution, and maintains audit trails to prevent hallucinated content from reaching production.
  • The Governance Compliance Auditor produces audit reports, maintains evidence logs, and sets remediation timelines.
  • 30 capability tags provide structured metadata for traceability, organized by capability, category, and supercategory.

Pillar 4: Scale

FCC is designed to scale from a pilot program to an enterprise documentation operation without architectural changes. The same framework that runs 5 personas in a base workflow can run 102+ personas in a complete workflow with champion orchestration.

What this means in practice:

  • Pilot (5 personas): Core personas handle the fundamental Find-Create-Critique cycle. No integration or governance overhead.
  • Team (20 personas): Integration personas add catalog indexing, taxonomy engineering, traceability, and blueprint validation. Governance personas add compliance, privacy, and anti-hallucination controls.
  • Enterprise (102+ personas): Champion personas orchestrate teams of base personas, producing unified research packages, orchestrated blueprint packages, multi-channel user experience deliverables, and validated operational packages.
  • The PersonaRegistry supports merging, so teams can define custom personas and merge them with the standard 102.
  • Workflow graphs are data (JSON), not code. New workflows can be defined without modifying the framework.

The Alternative

Without a framework like FCC, organizations that adopt AI for documentation typically encounter:

  • Inconsistency across documents because different prompts produce different styles, structures, and terminology.
  • Quality variance because there is no systematic review process -- output goes directly from generation to publication.
  • Governance gaps because compliance, privacy, and accuracy checks are performed manually (if at all) on AI-generated content.
  • Scaling limits because adding more documents means adding more manual coordination, negating the speed advantage of AI generation.

FCC addresses each of these failure modes structurally, not procedurally.

Next Steps