John McIntyre
Founder, HotPMO
Every month I trawl through the research papers, reports, and podcasts so you don’t have to. Here’s what caught my eye this month.
A strong month for the PMO’s reputation, and a sobering one for anyone delivering big infrastructure. PMI put out fresh evidence that having a PMO actually helps, two pieces landed on how AI adoption goes wrong when you treat it as a tool drop rather than an organisational change, and a pair of runaway railways offered a lesson in what happens when you build before you have finished thinking. I finish with a paper asking where projects come from in the first place, and a packed June diary.
Is your PMO an overhead or an insurance policy?
PMI’s Pulse of the Profession for 2026 turned its attention to complexity this year, and gave the PMO its own page of stats. Organisations with a PMO were more likely to rate themselves very or extremely successful at managing project complexity, 63% against 57% without one, and better at risk mitigation, 40% against 35%.
When I first saw that 63 against 57, I was a bit nervous. I would have expected a bigger gap. Then I read on and relaxed, because every single measure came out ahead for organisations with a PMO. Nothing pointed the other way. PMI goes as far as saying the PMO is the mechanism through which its M.O.R.E. mindset gets embedded across an organisation, which is quite something coming from PMI.
The framing I would take from it is that a PMO is a risk-reduction play. Set the cost of running one against the cost of the projects it oversees, the software, the people and the disruption to the business, and it is a tiny investment for a real reduction in risk. This is exactly the worry we keep hearing at the PMO meetups and on the PMO HotHouse, where people tell us their PMO is treated as an overhead. If that is you, this report is ammunition. Walk into the conversation with the numbers and the risk-reduction frame, and argue that you are saving the organisation money rather than adding to its costs.
AI adoption is an alignment problem with a human bill
Two pieces this month sit well together. The first, a systematic review by Huzooree in the Project Management Journal, pulls together 35 studies on generative AI in project management. Its argument is that the organisations getting value align three levels at once, strategy, process and people, instead of handing tools to individuals and hoping. What it wants is for AI to be treated as organisational design, a change you make across the whole organisation rather than a tool you install. That matches what I flagged last month: individuals feel faster, and the level above never sees the uplift.
The second, in Harvard Business Review, puts a name to the human side: psychological debt. Push AI at people and it can quietly erode a set of things that matter. The habit of thinking a problem through before reaching for the tool. The sense of your own competence, as tasks that used to take hours arrive in seconds. Everyday contact with colleagues, to the point where some universities are now spending real money to get students talking to each other again. Even credibility, because people use AI themselves and still take a dim view of others doing the same.
Two findings are worth holding onto. The debt falls hardest on the occasional users, the people who dip in and out, more than on the heavy daily users who seem quite happy. And it falls hardest on juniors. So if you are rolling AI out, those are the two groups to bring along with real care. One thing should raise alarm bells. If your organisation is treating an AI rollout as just another tool going in, that is your cue to challenge it, because this has to work at every level of the organisation or it works at none.
A check to run this month: for your current AI rollout, ask who owns the alignment across strategy, process and people, and ask which of those debts your occasional users and your juniors are quietly carrying. If nobody can answer the first question, you have found the PMO’s next job.
Two railways, one lesson
It was a big month for HS2. The Transport Secretary gave the six-monthly update to Parliament, and the Lovegrove Review landed the same day, this one looking at what the earlier James Stewart Review means for the civil service.
The cost story is by now familiar and grim. The estimate has climbed from £35bn to £45bn back in 2019 to somewhere between £87.7bn and £102.7bn to finish in today’s prices, or £70.9bn to £82.2bn if you hold prices at 2019 to compare like with like. Around £44bn has already been spent. And that is after cutting the top speed from 360 to 320 kilometres per hour to save a billion or two. Only about a third of the rise is inflation; the rest is scope that was missed and delivery that went wrong.
Two railways, both novel, both further from the platform than planned.
Two things stood out for me. The Lovegrove Review says large projects like this must have an internal audit function that is properly resourced and capable, and it takes HS2 to task for leaning on other government departments to do that job. If your PMO does assurance and governance, that is a line worth quoting upwards. The other is the pattern the reviews keep returning to: a focus on keeping things moving instead of spending the time up front to get the design right. Plan slow, build fast, and they did the reverse, with spades in the ground before the design was mature.
Then Stuttgart 21, which I put alongside it. Germany’s first fully digitised rail node, a brand new and largely unproven digital signalling system at enormous scale, on a similar if smaller journey: from about €2.5bn in the 1990s to over €11bn now, under construction since 2010, originally due in 2019, and as things stand with no firm opening date at all. Both are novel projects, and novelty carries risk. We accept that things go wrong on truly new work.
So a few questions for your own portfolio. Where are we starting to build before we have finished designing? Where we have something new and shiny, is there a real advantage to going bespoke, or would tried and tested do the job with far less risk? And the uncomfortable one: is your PMO known for challenging the spec, or for keeping things moving? Because in both these cases, moving fast before the thinking was done is exactly what cost the money.
Are we even doing the right projects?
Not everything this month was about AI, which makes a pleasant change. Söderlund and Locatelli have written an editorial in Project Management Journal called ‘Where Do Projects Come From?’, and I like the direction of it. Their point is that as a profession we pour our energy into doing projects right, and spend far less on whether we are doing the right projects at all, and on how projects get shaped and chosen before anyone starts delivering.
It is a call for more research rather than a set of findings, so there are no hard numbers to take away. There is a useful idea in it though: five lenses to look at an emerging project through.
- Politics: projects as arenas of power, coalitions and competing interests.
- Psychology: the decision biases, the planning fallacy and the value of an outside view.
- Marketing: shaping the market and the conditions for a project, rather than taking the market as given.
- Strategy: projects as one of the ways an organisation’s strategy actually forms, working in both directions.
- Design: creativity, prototyping and visualisation to bring an idea to life early.
Söderlund, J. and Locatelli, G. (2026), ‘Where Do Projects Come From? Project Shaping as a Foundational Domain of Project Studies’, Project Management Journal, Vol. 57(3), pp. 279–295.
The lens I would single out is psychology, and the planning fallacy in particular, which is exactly what we have just seen twice over on the railways: the confident belief that this time we have estimated it right. The pattern to watch for is escalation of commitment, where an idea gets backed at level after level until it is the emperor with no clothes and nobody feels able to stop it. A PMO is very well placed to be the one that points at the obvious project that is not fully dressed. Pick one shaping-stage idea in your pipeline this month, run it past the five lenses, and use a quick prototype to make it tangible early, so the problems surface before you commit real money.
A good day for the PMO
A quick word on International PMO Day, which fell on 12 May. There was a lot going on across London, and it was good to see so many PMO teams marking it. One that stood out came from North West Ambulance Service, who joined us on the show to talk about their day. They put the focus on their people, and one of their team, Matthew, had just won a Rising Star award for building their in-house project portfolio management tool. They have also contributed to the House of PMO’s Blue Light report on PMOs across the emergency services, which is worth a read if you work anywhere near blue-light delivery.
June: worth leaving your desk for
2 June: Laura Barnard’s free virtual session The IMPACT Engine: Helping PMOs Become Seen, Trusted and Valued, at 5pm.
10 June: PMO Flashmob London, a mid-year check-in aboard the Tattershall Castle on Victoria Embankment, from 5:30pm. I will be there, so do come and say hello.
17 to 19 June: the PMO Conference London, with PMOLearn on the 17th, the conference itself on the 18th (Alexander Budzier keynoting), and Laura Barnard’s IMPACT Engine Fundamentals workshop on the 19th.
Looking a little further ahead, the new Practical AI Skills for the PMO course, whose pilot sold out, runs again over the summer.
This roundup is based on my monthly PMO News segment from the PMO HotHouse, run by the House of PMO. If you’re not already part of that community, it’s well worth a look.
Building a challenge function into a PMO, the kind that questions a spec before the spades go in and asks whether a project is even the right one, rarely happens by accident. If that is the direction you want for yours, we would be glad to help.
