AI in Project Management: Moving from "What If" to "What Now"

By Emanuela Giangregorio

I was recently invited to contribute to APMG International’s series on artificial intelligence and project delivery. My article, Practical Examples of AI in Project Management, follows Antonio Nieto-Rodriguez’s piece on the future of the profession. Where Antonio explores where AI may take us, this article focuses on where it is delivering value today, and how to govern it well.

Here is a flavour of what the article covers.

The Adoption Gap

The Association for Project Management’s 2026 research found that over a quarter of UK project professionals already have AI fully embedded in their day-to-day work. That figure will only climb. Meanwhile, the inaugural Pulse of AI Governance report revealed that 77% of respondents have no structured, standardised approach to governing how AI is used in their projects.

Adoption is racing ahead of governance, and the governance gap is not closing fast enough.

Where AI Earns its Keep

In the article I walk through six application areas with concrete scenarios: planning and scheduling, risk management, monitoring and reporting, administrative automation, stakeholder engagement, and lessons learned. A few observations stand out for me.

Risk management may be the most compelling use case. AI excels at spotting patterns that individuals inside a project simply cannot see. Scattered references to delayed approvals, resourcing pinch points and unresolved design queries look trivial in isolation. Viewed together, they signal a delivery risk brewing.
Lessons learned finally has a fighting chance. Most organisations collect lessons; almost none apply them. AI turns a dormant repository into something a project team can actually interrogate in plain language. Institutional memory becomes retrievable rather than archived and forgotten.
The professional contribution does not shrink. A programme manager who once spent a full day compiling a consolidated report across ten workstreams can now review an AI-generated summary, dig into the two flagged as at risk, and spend the recovered hours talking to stakeholders. The analysis compresses. The judgement does not.

The Discipline is the Differentiator

The article closes with practical habits that separate effective AI use from careless AI use: starting with lower-risk tasks, learning proper prompting, keeping well-organised project records, validating outputs, and maintaining the domain expertise needed to spot when an AI output is wrong. None of these is technically demanding. All of them require deliberate habit.

And underpinning everything sits governance. One persistent misconception is that every AI use case needs its own governance regime. It does not. Sound AI governance is largely horizontal, which is precisely the thinking behind the AI Project Governance Framework: three principles (human-centricity, transparency and adaptability) applied consistently across the project life cycle, supported by a capability maturity assessment across four pillars.

Practitioners often assume governance slows adoption. The evidence points the other way. Teams with a clear framework, methods and templates adopt AI faster and realise the benefits sooner.