The New PMI AI Standard: What It Means for Project Managers in Manufacturing

In June 2026, the Project Management Institute published The Standard for Artificial Intelligence in Portfolio, Program, and Project Management, the first ANSI-approved global standard for using AI in professional project work. For a profession that has been absorbing AI tools faster than it has been governing them, this is a significant moment. It gives project managers a framework that did not exist before, and it raises a set of questions that manufacturing and NPI teams in particular should be thinking about.

This post explains what the standard says, what its eight principles and five performance domains actually mean in plain English, and what the implications are for project teams in regulated manufacturing and med-tech. It is not a substitute for reading the standard itself, but it is a starting point for thinking through how it applies to your projects.

Why This Standard Matters Now

AI tools are already in most project environments. Scheduling assistants, risk analysis tools, resource optimisation algorithms, meeting summarisation, and automated progress reporting are all in use by project teams right now. What most teams lack is a clear answer to the question: who is accountable when an AI-assisted decision turns out to be wrong?

That is the gap the standard fills. It does not tell you which AI tools to use. It tells you how to govern their use, how to maintain human accountability throughout, and how to manage the risks that are specific to AI in project work. For teams in regulated environments, where decisions have audit trails and every risk has an owner, this maps directly onto governance structures you should already have in place.

The Eight Guiding Principles

The standard is built on eight principles. They are technology-agnostic, meaning they apply regardless of which AI tools your team uses or how the tools evolve. They are also framed as principles, not rules, which means they require judgment to apply. Here is what each one means in practice:

The Five Performance Domains

The standard organises AI activity in project work across five performance domains. Each domain describes an area where AI can be applied and specifies how governance and human oversight should work within it.

What This Means Specifically for NPI and Capital Project Teams

Manufacturing and med-tech project teams work in environments where decisions carry regulatory consequences. A schedule decision that accelerates validation timelines has quality implications. A resource allocation decision that moves engineers off a design freeze has traceability implications. AI tools that influence those decisions need to be governed accordingly.

The three areas where the PMI standard is most immediately relevant to regulated manufacturing teams are:

1. Human-in-the-Loop at Phase Gates

The standard requires human-in-the-loop oversight at every stage where AI influences a significant decision. In practice, this means that if an AI tool produces a go/no-go recommendation at a phase gate, the recommendation must be reviewed and explicitly approved or rejected by a named decision-maker before any action is taken. The AI supports the decision. It does not make it.

2. AI-Specific Risks on the Risk Register

The standard is clear that AI introduces a category of risk that is distinct from standard project risk. Model accuracy, training data quality, and the risk of teams over-relying on AI outputs are all risks that should appear on the project risk register with owners and mitigation actions. This is additive to existing risk management practice, not a replacement for it. If your team is using AI tools on an active NPI programme, add an AI risk section to your register now.

3. Data Quality as a Project Governance Concern

The standard's data quality principle has direct implications for project teams in regulated environments. If an AI scheduling or resource tool draws on historical project data, that data must be accurate and representative. Using an AI tool trained on completed projects from a different product family or a different manufacturing context introduces a systematic bias that can be invisible until a deadline is missed. Data quality governance is now a project governance concern, not just an IT concern.

How This Relates to the EU AI Act

The PMI standard references the EU AI Act directly. For project teams in EU-based manufacturing companies, the standard provides a practical governance framework that supports compliance with the EU AI Act's requirements for high-risk AI applications. If your organisation is using AI tools that influence decisions in regulated product development, for example AI-assisted validation planning, design risk analysis or regulatory submission preparation, the PMI framework gives you a structured way to document oversight, maintain human control, and manage AI-related risk alongside existing quality systems.

This is not a compliance checklist. It is a principles-based framework that requires you to make judgements about your specific context. The EU AI Act compliance question is a separate and more complex one. But organisations that have adopted the PMI AI Standard's governance principles will be in a substantially better position when AI Act audits begin.

Where to Start if Your Team Is Already Using AI Tools

Most project teams are already using AI in some form, even informally. If that describes your team, the most useful starting points from the PMI standard are these three:

1. Inventory what AI tools your team is currently using and which project decisions they influence. This is the prerequisite for everything else.
2. Add a named human reviewer to every AI-generated recommendation before it is acted on. This can be done immediately, without new systems or process redesign.
3. Add AI-specific risks to your next project risk register. Model accuracy, data quality, and over-reliance are the three to start with.

If your team is not yet using AI tools in project work and wants to start, the PMI standard is the governance foundation to build from. Read the standard itself (ISBN: 9781628258912), then map your project governance framework against its five performance domains before introducing any AI tooling.

The broader project governance infrastructure that the PMI AI Standard assumes you have in place, phase gates, risk registers, formal change control, and named decision-makers, is described in our complete ISO 21502 guide for manufacturing PMOs. If those foundations are not in place, get them in place first. AI governance sitting on an ungoverned project environment helps no one.