Every major technological shift has changed not just the tools we use, but also the way work itself is organised. The Industrial Revolution transformed production by combining machinery with new processes, new skills and entirely new operating models. That transformation unfolded over decades. The internet age did something similar, reshaping how organisations communicate, access information, serve customers and coordinate work across boundaries. Its impact was felt over years rather than decades.
In both cases, the technology mattered enormously, but the real gains came from changing the systems, behaviours and ways of working around it. AI is the next significant shift and, while the technology may be different, the underlying lesson is familiar: technology alone does not create transformation. What is different this time is the pace. This shift is being measured in months.
While the adoption of AI is currently the hottest topic in town, it’s worth stopping to consider the fact that AI is ‘just’ technology. And like all technology it will require people and processes in order to change how we do things and importantly, to improve how we do things. AI tools ae unlikely to fill all of the gaps between poor process, unclear accountability and poor judgement. Processes and tools without capable people are unlikely to lead to much of a productivity gain, similarly capable people without the right tools and processes typically leads to inefficiency.
AI tools may offer significant advantages over the technologies that have gone before, but the central point remains the same. People still need to define how these tools are implemented, monitor how they are used and exercise judgement over the outputs they produce.
The organisations that gain the most from AI will be those that understand which decisions require human judgement, where that judgement needs to sit and where technology can genuinely accelerate or improve the work.
AI adoption shouldn’t be a race
There is considerable pressure within most large organisations to demonstrate that they are doing something with AI. In project and portfolio environments, that pressure is already translating into pilots, new tools, new use cases and increasingly ambitious expectations about productivity. The conversation is moving quickly from whether AI has a role to how fast it can be adopted.
But project delivery environments have never suffered from a shortage of technology. Over the past 20 years, organisations have invested heavily in PPM platforms, scheduling tools, dashboards, collaboration software, automated reporting and increasingly sophisticated data environments. These tools have undoubtedly improved visibility, speed and access to information. Yet many of the underlying problems remain stubbornly familiar: poor planning, unclear accountability, inconsistent governance, poor-quality data and a lack of sponsorship and management support for initiatives that become troubled.
That should give us pause. Technology can improve the way work is performed, but it cannot compensate for every weakness in the system around it. The real challenge is therefore not how quickly organisations can adopt AI. It is whether they understand the role that people, process and technology should each play in creating better project and portfolio outcomes.
We have been here before
The adoption of the internet provides a useful comparison. In the 1990s and early 2000s, organisations invested heavily in connectivity, enterprise systems and new digital tools, often with the expectation that the technology itself would unlock major productivity gains. In reality, the benefits were far more dependent on what organisations changed around the technology, including their operating models, skills and ways of working.
Blockbuster is a useful example. According to a Harvard Business School Case study, the company is often characterised as a business that simply failed to recognise the internet, but the reality is more nuanced. By the end of 2006, Blockbuster had built an online rental service with around 2.2 million subscribers, including approximately 2 million paying customers. It had also launched an integration between its online service and physical store network. Blockbuster did adopt the technology and was investing heavily in it.
The challenge was that the new technology was still being incorporated into an operating model built around thousands of physical stores. Netflix, by contrast, had developed its business around the online subscription model and continued to adapt that model as the technology evolved.
Project and portfolio management offers a smaller-scale version of the same lesson.
Over the past two decades, organisations have invested in many types of PPM specific tools. These technologies have undoubtedly improved access to information, reduced administration and increased visibility. What they have not done, on their own, is resolve the harder organisational issues that sit behind delivery performance.
The lesson is not that technology is ineffective. It is that technology cannot compensate for weaknesses in the system around it. In a well-run delivery environment, better tools can strengthen visibility, accelerate analysis and make good practice easier to apply consistently. In a poorly run environment, the same tools will still produce faster reporting, more data and more automation. But if the underlying data, processes or assumptions are weak, the outcomes will be too. You will simply get them faster.
Human judgement becomes more valuable, not less
As AI takes on more of the administrative and analytical work, the value of human capability is likely to move further up the chain rather than diminish.
Project and portfolio environments are full of activities that can be accelerated through technology: consolidating data, producing reports, identifying trends, drafting updates and testing scenarios. But the most important decisions rarely depend on information alone. They depend on an understanding of nuance, subject matter expertise and human judgement.
Someone still needs to challenge an unrealistic assumption, recognise when the data does not tell the full story, understand the politics around a decision, balance competing priorities and determine when an issue needs to be escalated. Someone needs to distinguish between a project that is genuinely under control and one that is simply reporting green. These are not usually administrative tasks, or at least they shouldn’t be. They require context, experience, curiosity and, at times, the confidence to deliver an uncomfortable message.
Rather than reducing the importance of project professionals, AI may make the distinction between different types of capability much clearer.
Roles that are heavily focused on collating information, administering process and producing routine outputs are likely to change significantly. Greater value will increasingly sit with people who can interpret information, challenge assumptions, exercise judgement and help leaders make better decisions.
The question for project professionals, therefore, is not whether AI will take over parts of the work. It undoubtedly will. The more important question is what we do with the capacity that creates and whether we are developing the judgement and capability needed to use it well.
Where AI genuinely changes the equation
Where AI becomes genuinely interesting is not in producing another project report faster. It is in its ability to work across volumes of information that would previously have been impractical or very time consuming for a person to interrogate routinely.
In project assurance, for example, one of the challenges is rarely a lack of information. It is the opposite. A major program can generate schedules, risk registers, financial reports, steering committee papers, status reports, action logs and hundreds of pages of supporting material every month. The real effort is in connecting those sources. Identifying where they contradict one another and determining what deserves attention.
This is where AI has the potential to change the economics of the work. It can help interrogate large volumes of project information, compare narratives with underlying data, identify recurring themes and surface anomalies for a person to investigate further. It does not need to make the judgement to be valuable. Finding the five things that warrant human judgement within 5,000 pieces of information is valuable in itself.
There is already evidence of meaningful productivity gains from this type of augmentation.
Research by Brynjolfsson, Li and Raymond found that access to a generative AI assistant increased productivity among more than 5,000 customer service workers by an average of 14 per cent, with particularly significant gains among less experienced workers. The researchers found evidence that the technology was effectively helping disseminate the practices of higher-performing employees.
Research within the project profession is pointing in the same direction. PMI’s 2024 study of 500 project professionals found that heavy users of generative AI were substantially more likely to report improvements in productivity, problem-solving and effectiveness than lower adopters.
The real opportunity is therefore not to automate judgement, but to remove the effort that sits in front of it. If AI can reduce the time spent assembling information, identifying anomalies and surfacing areas that warrant attention, project professionals can spend more of their time on the work that creates value: interpreting what the information means, challenging assumptions and making better decisions. That is a fundamentally different proposition from simply doing the same work faster.
The PMO is best placed to help design a better delivery operating model
AI creates an opportunity for PMOs to do more than introduce another tool or automate an existing reporting process. The more important role is to help organisations rethink how project and portfolio delivery should work when technology can take on a greater share of the administrative and analytical load.
That means asking some more fundamental questions.
Which activities should be standardised? Which can be automated? Where are we getting bogged down in the quantity of data? Where is human judgement essential? What information do leaders actually need to make decisions? Which controls genuinely improve delivery and which simply add effort? Where are capability gaps creating risk that technology cannot solve?
High performing PMOs are well placed to lead this work because they sit across projects, portfolios, governance, reporting, assurance and capability. It supports the strategic link underpinning why PMOs exist in the first place. They can see where information is duplicated, where decisions are delayed, where processes are inconsistent and where people are spending time on activity that adds little value.
These are operating model questions, not delivery questions and not technology questions.
The opportunity is to move beyond being the custodian of processes, templates and tools and become a truly strategic function that actively improves how delivery operates. In an AI-enabled environment, the most valuable PMOs will be those that help organisations design a better combination of people, process and technology, with clearer accountabilities, better information and more effective decision-making.
The organisations that benefit most from AI will probably not be those that deploy it fastest. They will be those that redesign the work around it most intelligently.
In July 2027 Pledge Consulting will be hosting the 3rd Project Management Technology Showcase in Sydney. The aim of the event will be to bring together project delivery professionals, PMO, strategy leaders and technology vendors, to talk about and showcase the optimal mix of People, Processes and AI. For more information, see our website.https://hi.switchy.io/_3Da
References:
Brynjolfsson, E., Li, D. & Raymond, L.R. (2023), “Generative AI at Work”, NBER Working Paper 31161. The study examined 5,179 customer support agents and found a 14% average productivity increase, rising to 34% for novice and lower-skilled workers.
NBER paper
Project Management Institute (2024), “First Movers’ Advantage: The Immediate Benefits of Adopting Generative AI for Project Management”. PMI surveyed 500 project professionals and found that heavy GenAI adopters were more likely to report improvements in productivity, problem-solving and overall effectiveness.
Willy Shih, Stephen P. Kaufman and David Spinola, “Netflix”, Harvard Business School, Case No. 607-138, published 31 May 2007.