Every project manager I know has been handed the same advice for two years straight: “use AI to be more productive.” Useful, right? It is the productivity equivalent of telling someone to “just exercise more.” The question was never whether AI can help a PM. The question is where, exactly, it pays off, and where it costs you more time than it saves once you account for checking its work.
After putting AI through a lot of real project weeks, here is the honest map.
Start with the busywork, not the big calls
The instinct is to point AI at the hardest part of the job: the call on whether to cut scope, the read on whether a vendor is bluffing, the judgment about which risk is real. Resist it. Those are exactly the places AI is weakest and the stakes are highest.
Point it at the busywork instead. The reformatting, the first drafts, the summarizing, the “turn these twelve messy bullet points into something a stakeholder can read.” That work is low-stakes, high-volume, and it eats hours every week. AI is genuinely good at it, and a wrong answer costs you a quick edit rather than a blown decision.
Status reports are the killer use case
If you try one thing, make it this. You already have the raw material: stand-up notes, a few Slack threads, a list of what closed this week. Hand AI the raw material and your usual format, and ask for the draft.
It will not nail the tone on the first pass. It does not need to. It gets you from a blank page to eighty percent in about a minute, and the remaining twenty percent (the part that needs your judgment about what to highlight and what to soften) is the part only you can do anyway. You stop spending your Friday afternoon assembling facts and start spending it deciding what they mean.
Planning: a thinking partner, not an oracle
AI is a good planning partner precisely because it is tireless and a little bit dumb in a useful way. Describe a piece of work and ask it to break the task down, and it will produce a list that is roughly right and obviously incomplete. That is the point. Arguing with a draft is far faster than generating one from nothing.
Ask it: what am I forgetting, what usually goes wrong with work like this, what would a skeptical stakeholder ask. You are not outsourcing the plan. You are using a fast first draft to surface the gaps your own experience can then fill.
Risk reviews and the pre-mortem
The pre-mortem (“assume this project failed, now explain why”) is one of the most valuable exercises in project management and one of the hardest to run, because the people in the room are too close to the work to imagine it failing. AI has no such attachment.
Feed it the project and ask it to argue, hard, that the thing will miss its date. You will get the obvious risks you already knew, plus two or three you had filed under “probably fine.” Those two or three are the entire return on the exercise. Then you do the part that matters: deciding which ones are worth acting on.
Where AI gets in the way
It is not all upside, and pretending otherwise is how people end up trusting it past its limits.
- Anything that needs a source of truth. AI will state a date, a number, or a dependency with total confidence and be flatly wrong. For anything load-bearing, it drafts; your system of record decides.
- Reading the room. It does not know that this stakeholder is nervous about budget or that this engineer has been burned before. The political and human layer of the job is still entirely yours.
- Net-new structure when you do not know the answer yourself. If you cannot tell whether the output is good, you should not be shipping it. AI multiplies your judgment; it does not replace it.
The skill that actually matters
Here is the part the hype skips. The PMs who get real leverage out of AI are not the ones with the cleverest prompts. They are the ones with the judgment to know, instantly, whether the output is any good. That judgment does not come from a tool. It comes from having run enough real projects to feel when something is off.
Which is the whole reason the training and the apps we are building start from thirty years of practice rather than from the latest model. AI is an amplifier. It makes a good PM faster and a careless one more confidently wrong. The work, as ever, is in being the former.
If you want the full version of this (the prompts, the workflows, the worked examples), it is the heart of the AI-for-PMs eBook and video series. Drop your email on the waitlist and you will be first to get it.