A methodology pipeline. Not a session.

Any modern AI system can produce a Charter from a brief. The output will read well. It will use PMBOK vocabulary. It may even cover the standard 13 charter sections.

A pipeline does something different.

A pipeline produces 39 artifacts that are internally consistent - the risk register and the charter reference the same assumptions, the WBS aligns with the schedule, the stakeholder register matches the communications plan, the budget contingencies reconcile with risk response strategies. A pipeline enforces this consistency through architecture. A one-shot generation enforces it through attention - and attention degrades across 39 artifacts in a long session.

A pipeline is reproducible. The same brief produces the same artifact set, every time. A pipeline can be audited because the output is deterministic. A one-shot generation produces a different set each run.

A pipeline persists organizational memory. The next project starts from the last project’s knowledge base, not from zero. A one-shot generation forgets between sessions.

A pipeline produces executive-level synthesis - a composite grade against PMBOK Performance Domains, a Kerzner Maturity Assessment, a Critical Findings analysis, a prioritized Decisions and Actions list. Not just the artifacts, but the read of the artifacts.

A pipeline is a procurement-grade workflow. A one-shot generation is a session.

This is the difference. It is structural, not cosmetic.

For the full comparison including the system prompt we used in our own parity test, see When PMAC, When Cowork, When Both.

Every PMAC run produces 43 artifacts total. 39 ship to you as the customer-facing output set. 4 stay on-platform as a process / output / data audit trail - what we validate every delivery against.

Chain Configuration

Process

Run-time configuration: model parameters, token usage, agent execution sequence

Co-Pilot Analysis (mid-chain)

Output

Agent 5 quality assessment of upstream agents' outputs

Finalized Knowledge Base (raw)

Data

Unfiltered cross-agent knowledge base scaffold; customer-facing KB ships through our audience-filtered variant

Co-Pilot Analysis (end-chain)

Output

Agent 6 end-of-chain quality assessment of the full sequence

Three audit dimensions - process, output, data - run on every project. This is what separates PMAC from “AI generated my project plan” tools: we ship audited output, not raw model generations.

The pipeline’s quality gate operationalizes PMBOK’s eight Performance Domains as the rubric for evaluating every artifact produced. Each domain decomposes into observable sub-criteria. Weights derive from PMI Pulse of the Profession 2024 failure-cause analysis.

Stakeholders

16%

alignment, engagement, communication

Uncertainty

16%

risk identification, response, monitoring

Planning

16%

WBS, schedule, dependencies, estimates

Project Work

12%

resource management, knowledge transfer

Delivery

10%

scope, quality, value realization

Measurement

10%

KPIs, earned value, performance reporting

Team

10%

composition, leadership, governance

Development Approach

10%

lifecycle fit, methodology selection

Composite grade thresholds: A (90% and above), B (80 to 89.9%), C (70 to 79.9%), D (below 70%, triggers internal quality fallback).

PMAC is not certified by the Project Management Institute (PMI) or licensed by the International Institute for Learning (IIL). Methodology fidelity is maintained through internal audit and external peer review with practicing senior project managers. That distinction is deliberate.

Methodological honesty. Certification implies an institutional pathway PMAC does not follow. Benchmark-against is accurate.

Procurement defensibility. Enterprise buyers running formal procurement review prefer honest disclosure over claimed certification that cannot be verified.

Methodology evolution. PMBOK 7 and Kerzner PMMM are reference frameworks, not version-locked specifications. Benchmark-against allows methodology refinement without breaking certification claims.