There’s a lot being said about AI in PMOs at the moment.

Some of it makes sense. Some of it… not so much. It can feel like everyone agrees it’s important, but no one quite agrees on what that actually looks like day to day.

And that’s usually where people get stuck. Not on the tools. On the “what do I actually do with this on Monday morning” question.

The Practical AI Skills for the PMO course seems to start from that point. It doesn’t try to position AI as something completely new. It’s more like, here’s the work you’re already doing. Now let’s look at where AI quietly fits into it.

Which, when you think about it, is probably a more useful way to approach it.

Reporting and insight, where most people first notice a difference

This tends to be the entry point.

Not because it’s the most strategic use of AI, but because it’s the most visible. And, if we’re honest, one of the most time-consuming parts of PMO work.

Pulling together updates, chasing inputs, rewriting things so they sound consistent. It’s not difficult work, but it is repetitive. And it adds up.

AI can help by drafting summaries, pulling together multiple updates, even highlighting where things don’t quite line up. I’ve seen it take a fairly messy set of inputs and turn it into something that reads almost like a finished report. Not quite, but close enough that you’re editing rather than building from scratch.

Which changes the dynamic a bit. You spend less time assembling information and a bit more time thinking about what it actually means.

That shift sounds small. It probably isn’t.

Governance and assurance, or moving past going through the motions

This is one of those areas that looks strong on the surface.

Frameworks, checklists, stage gates. Everything appears in place. And yet sometimes it still feels like you’re proving that something exists rather than proving that it works.

AI can generate governance content quite easily. That part is straightforward.

What’s more interesting is using it to challenge that content. Asking it to act almost like an external reviewer. Where are the gaps, what assumptions are being made, what feels too neat.

It also highlights something that comes up a lot. Good governance is not just about having the right artefacts. It is about whether those artefacts actually reflect reality. And that is harder than it sounds .

Risks, issues and dependencies, or trying to see the whole picture

Most PMOs already know this is difficult.

Risks sit in one place, issues in another, dependencies often in people’s heads or buried in conversations. You end up piecing things together manually, and usually a bit too late.

AI is quite good at pulling those threads together. It can scan different sources, group related risks, surface patterns. Things that are technically there already, just not visible.

I think this is where it starts to feel less like automation and more like augmentation. It is not replacing the analysis. It is just making the connections easier to see.

Although, having said that, it can sometimes surface more than you actually want to deal with. Which is a slightly different problem.

Planning and scheduling, and the question of whether the plan actually works

AI is very good at producing plans.

Give it a brief and it will return something structured, logical, well laid out. It looks right. And in many cases, it’s a really useful starting point.

Because what it produces is often a “standard” plan. Clean, balanced, and based on recognised patterns. Which is helpful… although it doesn’t always reflect the specific constraints you’re working with.

So the more valuable use case isn’t just creating the plan. It’s building on it.

You can use AI to explore timelines, introduce constraints, and test different scenarios. It’s particularly good at asking “what happens if this slips?” or “what changes if this is delayed?” questions that are easy to overlook when you’re close to the work.

It adds another layer to planning. Not replacing judgement, but supporting it. Helping plans feel a bit more grounded, or at least more thought through before they’re put into action.

Resource and portfolio decisions, where things rarely line up neatly

This is where conversations tend to get complicated.

There is usually more work than capacity, more priorities than can realistically be delivered, and a fair amount of negotiation happening in the background.

AI can help model different scenarios. Not perfectly, but quickly.

What happens if this project is delayed, what happens if that resource is reallocated, what is the least disruptive option across the portfolio.

It does not remove the need for judgement. In some ways it highlights it. Because you still have to decide which trade-off you are willing to accept.

But it does mean you are working with options, rather than just constraints.

Meetings, documentation and the quieter parts of PMO work

There is a lot of effort here that tends to go unnoticed.

Capturing discussions, writing up notes, tracking actions. It is not complex work, but it is constant.

AI can take a fair amount of that on. Transcribing meetings, summarising conversations, extracting decisions and actions.

I used to think this was more of a convenience than anything else. It is also about accuracy. And consistency. Things that tend to slip when people are busy.

So it ends up being more useful than expected.

Decision support and scenario thinking, where things start to shift

This is probably where the role of the PMO begins to change slightly.

AI can help model scenarios. Budget changes, delays, different combinations of projects. It gives you a way to explore possibilities without spending days building models.

The outputs are not definitive. They should not be treated as such.

But they do provide a starting point. Something to react to, challenge, refine. Which is often what is needed in senior conversations.

Not certainty. Just a better-informed view.

Creativity and PMO development, which feels less obvious but comes up anyway

This one is a bit harder to pin down.

Using AI for creativity can feel a little abstract at first. It’s not always the obvious place people start. But in practice, it becomes useful quite quickly, especially when you’re trying to design or improve something.

That might be a new PMO service, a process, a workshop, or even how you approach training.

AI is good at getting things moving. It can generate ideas, suggest different approaches, and help you see options you might not have considered straight away. Some ideas will land better than others, of course, but that’s kind of the point. It gives you something to react to.

It’s also helpful when it comes to how you present information. AI can turn content into reports, visuals, simple infographics, or structured summaries, making it easier to communicate ideas clearly to different audiences.

So it’s not just about creativity in the abstract sense. It’s about momentum. Instead of starting with a blank page, you start with something. And from there, it’s much easier to shape, refine, and build something that works.

So what does all of this add up to

There’s a pattern running through all of this, although it’s not always obvious at first. AI isn’t taking over PMO work. It’s taking over parts of it. The drafting, the aggregation, the initial analysis. The more mechanical tasks that take time, but not necessarily judgement.

What remains is the harder part. Judgement, challenge, interpretation. Knowing when something doesn’t quite add up.

In theory, that’s always been where the PMO adds value. In practice, though, a lot of time gets pulled into the mechanics. Reports, logs, updates. AI shifts that balance. Not completely, but enough to create space for more meaningful work.

That’s really the focus of the Practical AI Skills for the PMO course.

It’s built around realistic PMO scenarios and guided exercises, showing how AI fits into everyday work. There’s a strong emphasis on using AI safely and ethically, alongside practical skills like prompting, guiding outputs, handling data properly, and identifying where tasks can be automated or improved.

It’s less about the tools themselves, and more about how to use them effectively in a PMO context.