The hardest part of using AI in projects isn’t the technology. It’s knowing what it can really do for you, and whether you can trust what it gives back. Most project and change leaders were never taught either, so they stick to the safe, simple tasks and leave the real value untouched.
Ask a PMO what AI has changed and the answer is usually the same. Meeting notes write themselves. Status reports are put together in seconds. Some teams have gone further and bought tools that score risks and forecast delivery dates.
All of that is real and valuable. But none of it is transformation. The steering committee still meets monthly, the stage gate still sits where it always did, the baseline still gets defended. AI arrived, and we gave it the paperwork. That is the smallest thing it can do.
The reason to care about AI in projects is not speed. It is the possibility of seeing what we currently cannot see.
Projects rarely fail without warning. Looking back, the signs were there for months. A team that stopped reporting issues. A dependency slipping by a few days repeatedly. An approval cycle taking quietly longer. Nobody connected them, because connecting hundreds of small signals across a portfolio was never realistic by hand. But that is exactly the kind of work AI is good at.
So why do so few teams start leveraging that capability?
Not because they picked the wrong tool. Because seeing further depends on what AI is looking at, and that part is ours.
Everything AI tells you about your project comes from data somebody chose. Which signals get captured. How they are recorded. Who touches them before they arrive. Whether a green status means the work is fine or means nobody wanted to explain an amber status. The model does not know the difference. It works with what it is given.
This is why the skill that matters is not prompting or tool selection. It is understanding the input. Knowing what an output rests on, where it probably breaks, and how much weight it can carry for the decision in front of you.
None of that requires data science. It requires the habit of asking the right questions before acting on a number.
I write about what data and AI can genuinely do for projects, and the literacy it takes to tell the difference between an output worth acting on and one that only looks convincing.