The Canopy Method

Designed against the PMI Standard for AI in PPPM

PMI wrote the standard.
Canopy runs it.

PMI’s Standard for Artificial Intelligence in Portfolio, Program, and Project Management defines the principles for adopting AI responsibly. The Canopy Method is the executable system that instantiates the human-oversight, governance, and decision-accountability principles in shipped, CI-enforced code — not a framework deck.

Two layers, not two rivals

Governance is a document until something enforces it. Canopy is that something — a complement to the standard, not a competitor to it.

PMI governs

  • Voluntary consensus guidance — a document
  • Tells human managers how to adopt AI as a supervised tool
  • Sits on top of whatever method you already run
  • Eight principles, unweighted; explicitly non-enforceable

Canopy runs

  • Primitives enforced in shipped code, checked in CI
  • A runtime where humans and AI agents do the work under one audited vocabulary
  • Replaces the method — continuous Flow, no ceremonies
  • A small enforced contract your build actually verifies

Principle by principle — honestly

Where Canopy operationalizes a PMI principle in shipped code, and where it deliberately declines. No principle is “fully covered” — that would not be true.

Risk (human-in-the-loop)

Operationalizes

Ships: A first-class approval gate, capability-scoped human-vs-agent identity, and every override recorded in an immutable audit trail with its reason and approver.

Does not: Does not detect AI bias, hallucination, or drift. The gate is enabled by policy, not on by default.

Governance & Compliance

Operationalizes

Ships: Approval gates, an immutable access-controlled audit trail, and a human-vs-agent authority model — surfaced in a live Approval Gate Queue and a weighted Risk Heat-Map.

Does not: Not regulatory compliance. It produces the audit substrate that feeds compliance; it does not verify against any law.

Ethics & Professional Responsibility

Operationalizes

Ships: Every action carries a human-or-agent identity; consequential moves can be held for a recorded human approval; who-decided-what stays traceable and reviewable.

Does not: No fairness, explainability, or privacy framework. Accountability is operational; ethics modeling is not.

Strategic Value

Partial

Ships: Every unit of work is an Outcome — a result that delivers value, not a unit of activity — with real lead-time and throughput so value ships every week.

Does not: Delivery is measured; business value and ROI are not quantified.

Stakeholder Engagement

Partial

Ships: Audience-tailored briefings — technical, business, and regulatory lenses over one source of truth — plus a live customer Portal and an AI-drafted, human-edited Weekly Pulse.

Does not: No stakeholder segmentation or AI-adoption change management.

Optimization & Innovation

Partial

Ships: AI generates your Signal feed, attention alerts, and a draft Pulse — a human decides — while lead-time, throughput, and decision-velocity trends make efficiency measurable.

Does not: No model optimization or self-learning loop.

Data Quality

Partial

Ships: Context is a first-class, living, version-traceable input — PMI's input-for-impact principle, made an entity — backed by an immutable audit trail.

Does not: No data-governance framework or model-pipeline lineage.

People & Culture

Declines

Ships: A team-adoption tool onboards people to the method's vocabulary, and the human/agent split makes roles explicit.

Does not: No leadership, literacy, or change-management framework. A deliberate blind spot.

What that looks like in the product

Human-in-the-loop, in code

AI-agent actions carry a distinct, capability-scoped identity. Consequential Outcomes can be held for explicit human approval. Every override is permanently recorded with its reason and the approver it bypassed.

Governance you can see

Approval gates, an immutable audit trail, and a human-vs-agent authority model are enforced in the system and checked by seven CI gates — surfaced in a live Approval Gate Queue and a weighted Risk Heat-Map.

Value ships continuously

Every unit of work is an Outcome — a result, not an activity. Canopy tracks real throughput and lead time with a four-week trend, so teams watch value ship every week instead of at the end of a long cycle.

Decisions are accountable

Decisions are a first-class, measured primitive and the audit trail is structurally immutable, so AI accountability is operational, not aspirational.

What Canopy is not

The standard is broad on purpose; Canopy is narrow and enforced on purpose. Canopy is the project-execution runtime for AI governance. It is not:

  • An MLOps or data-governance platform
  • A detector of model bias, hallucination, drift, or explainability
  • Regulatory compliance, legal, or contract tooling
  • A change-management or AI-literacy program
  • Affiliated with, endorsed by, or certified by PMI

The complete picture

Canopy is one clan in a complete operating system

PMI organizes the world as portfolio → program → project, under governance. Nfinit Monkeys built that as a living, AI-native system — and Canopy is the execution method inside one tier of it. Every layer PMI describes is a real, running primitive.

Biome
Portfolio

The whole ecosystem of ventures.

District
Program

A coordinated line of business.

ClanCanopy
Project

A product or initiative team — where the Canopy Method runs.

Alpha
Accountable owner

The lead that owns a clan's outcomes.

Jungle
Governance & oversight

Council governance, approvals, and orchestration — the BOSS layer.

Mandrill
Reusable capability

Skill packs agents wield, published once and reused.

Swarm
The AI workforce

Parallel agents executing the work.

PMI wrote 297 pages on how to manage AI across portfolio, program, and project. Nfinit Monkeys built it — and runs on it. Canopy is the part you can adopt today; BOSS (the Business Optimization & Saving Suite) is the operating system behind it.

Complete for how we run our own work. You adopt the operating system piece by piece — Canopy is the entry point, not the whole obligation. Your legal, regulatory, and people duties stay yours; we give you the runtime and the governance to discharge them.

Canopy is the only post-Scrum methodology whose human-approval gate, capability-scoped agent identity, and immutable audit trail are shipped as executable, CI-enforced code — operationalizing the PMI Standard for AI’s principles of Risk, Governance, and Ethics at the project-execution layer.

The Canopy Method is an independent product of Nfinit Monkeys. It is not affiliated with, endorsed by, certified by, or sponsored by the Project Management Institute. PMI, PMBOK, and Disciplined Agile are trademarks of the Project Management Institute, Inc. References to PMI’s Standard for Artificial Intelligence in Portfolio, Program, and Project Management are for descriptive comparison only. Capability descriptions reflect shipped functionality as of publication.