The answer as to whether machine learning and artificial intelligence will replace various roles in the workplace is one contingent on several very big factors; most notably the nature of the work in question and the relative sophistication of AI in the coming years and decades.
For some aspects of project management there is the potential for AI to automate certain functions but in terms of making PM’s obsolete that’s another question. Let’s take a closer look at where AI and automation is coming on in leaps and bounds and where it might have a natural advantage over the human brain.
1. Automation Opportunities – Data Analytics
One of the major challenges for PMs is the enormous amount of data that modern projects produce. This has to be assimilated and analysed in order to produce meaningful reports for project sponsors and senior managers. There are times, dates, assignments, costs analysed in various ways, progress on deliverables, checklists for compliance, deadlines for reports and meetings, resource requirements and ever-changing risk assessments.
Add to these the fact that the business, political and international landscape is in flux, and it’s clear that no single project manager can possibly absorb and analyse all the data. Yet the PM is expected to take evidence-based decisions, based on this data.
So this would seem to be a prime example of big data that can have an AI rule set applied to it and which would then produce results that could reasonably qualify as business intelligence. This would then serve as the basis for informed decision making. It is quite feasible. However, it may not happen, for reasons that are inherent in the very nature of projects.
2. Problems with Cost Effectiveness for AI
The problem lies in the nature of projects themselves. The very reason that something is declared a project – it is a one-off piece of work with a defined scope and outputs – is also the reason that AI may struggle to be cost-effective in the project sphere.
Different projects produce data that is different in kind, as well as in quantity. It’s not worth driving a set of AI rules through the learning process when they have to be changed radically for each project. To set up an AI application is expensive – it only becomes cost effective if it can be used repeatedly to automate standard tasks. Unfortunately, standard tasks are not common on projects.
This is why the leading methodologies, such as the PRINCE2 project management methodology are not prescriptive. They tend to define a method with processes and boundaries that can be applied to all projects. There is no attempt to get involved in the detail – because the detail is always specific to a given project. So the economics of AI in project management remain to be proven.
3. PMs becoming more People Centred in their Approach
The other problem for AI is that its potential arrival in the project management sphere coincides with a move in the opposite direction as far as project managers are concerned. Project managers are becoming aware that they have drifted some way from a corporate culture that has been embracing transparency, diversity, accountability and initiative. This is particularly the case as millennials enter the job market after university.
The old “command and control” style PM is giving way to a more empathetic and people-centred manager who expects to explain the reasons for decisions, and build motivation and engagement in the project team. This is particularly important where the project is building cultural products such as TV, film and so on. Who is to say whether a TV series has delivered its brief? We’re certainly a long way from any AI system that can critically review creative work.
Without this, the AI system will be restricted to those roles that are currently carried out by the project administrators. Invoice processing, time tracking, meeting arrangements, updating plans and producing forecasts are not really project tasks. They are standard office administration tasks that are taking place in a project context.
These roles will soon be under threat from AI. This is because invoice processing, for example, is very similar in most organisations – so building an AI system to deal with it is cost effective. In other words the opposite scenario from the introduction of AI into project management.
4. The PM Role is Business Wide
The other reason that AI will struggle to take over project management is the scope of the PM role. Initially, AI systems will come into a business function – let’s say the accounts group, and will take over aspects of the work, probably working alongside human supervisors. It will be a very long time before AI is allowed to range across an entire business, networking independently in order to get work done.
Yet PMs routinely have to do this, working across an organisation to secure resources or sponsorship, or to engage the future users of the project’s deliverables. Successful PMs spend a surprising amount of their time exercising soft skills, such as negotiation and consensus building. So we may not see AI taking over PM roles until they develop an AI system that can take the Senior User to lunch!
About the Author: David Baker is marketing manager at PRINCE2 Training, who provide courses and certification in PRINCE2, Agile, Lean Six Sigma, ITIL, PMP, and Scrum project management methodologies. You can connect with David and PRINCE2 Training on Twitter, Facebook and LinkedIn.