Skip to content

We must seek to understand how AI can improve (not replace) the project profession

Added to your CPD log

View or edit this activity in your CPD log.

Go to My CPD
Only APM members have access to CPD features Become a member Already added to CPD log

View or edit this activity in your CPD log.

Go to My CPD
Added to your Saved Content Go to my Saved Content
Gettyimages 1142994736

There is no doubt about the impact of AI in many areas of business. Wherever there is lots of data about repeated events, AI can be used to find patterns, predict what comes next, diagnose problems and so on. This also applies to projects where the same activities are repeated many times.

However, for large and/or complex projects, or for ones which are full of ‘first of a kind’ situations, AI’s utility is less clear. Here, the role of the project professional – managing the people who deliver the project – will remain. Or will it? At the heart of this debate are several questions.

What are project managers’ competencies?

You might have thought that the answer has been documented clearly in bodies of knowledge. However, it may be appropriate to re‑evaluate competencies, particularly the distinction between behavioural and other competencies, but also to review them in light of how AI is developing, so that competencies can be classified according to whether AI replaces or supports the competencies or threatens their deployment.

This applies particularly to complex projects, where the relationship between competencies and project success is most critical.

What do project professionals actually do, and how do they do it?

What project professionals do is generally understood, although the time and effort spent on different activities varies between projects, roles, levels of seniority, etc. Research into the role of AI should avoid generalisations and instead focus on specific examples of the deployment of AI to replace or support project professional activity.

With which tasks can AI support project professionals to work better or faster?

A key issue is how human roles can be combined with AI, as opposed to being replaced by it. The idea is that AI can be used to improve predictions of outcomes of particular project actions, including enhanced risk analysis. However, there is not much reliable and deep public evidence of how project planning and delivery have been affected and, more importantly, what tools and techniques have been deployed in practice and how they should be developed and implemented.

What data development is required to ensure AI can be deployed?

AI thrives in the world of big data. If AI is to improve project management, it will need much more data from projects, perhaps even by‑the‑minute reporting of project status. Much of the extra data needed for AI to be deployed successfully is unstructured, e.g. project professionals’ opinions about risks, and may not even be captured now.

So, much effort will be needed for identifying and collecting many different sorts of data – structured and unstructured. There is no presumption that these new data sets will be perfect. The key is to identify and make use of them, learning through AI which data sets are useful, and in what forms, and where improving the quality of the data might bring returns.

How will AI be deployed to analyse and predict?

Data analysis may best be done by combining human and artificial intelligence, e.g. by humans initially identifying the meaning of data and then training the AI to generalise from these classifications – so‑called ‘supervised learning’.

Once all the data becomes analysable, the idea of a digital twin for a project comes into its own. Digital twins thrive in situations where high volumes of data are used to optimise management of technical artefacts (e.g. buildings, airliners). The question is whether, in projects that involve substantial behavioural change, the digital twin approach can be used to plan, model and manage delivery.

What will the benefits of the deployment of AI be?

The benefits of applying AI to project management are expected to include:

  • creation of a stronger and more widely shared basis for decision‑making;
  • increased rationality, especially via removing/reducing decision‑makers’ cognitive bias;
  • more accurate forecasting of project progress and completion;
  • increased speed of decision‑making;
  • improved identification of missing or imperfect data;
  • better incorporation of learning from experience; and
  • higher-quality management of projects and resulting higher success rates.

Our understanding of AI has developed greatly in the past few years, but we must dig deeper to understand how AI can improve project management, other than by substituting automated analysis for routine tasks. The main focus of our work should be on the most central element of project performance, the human factor. We believe more research into this is needed.

See the Autumn 2022 edition of Project journal for a longer version of this article

Read the new APM research on AI and projects:

This was co-written by Brett Parnell and Professor Merlin Stone.

Professor Merlin Stone is Principal at Merlin Stone Consulting.




Join the conversation!

Log in to post a comment, or create an account if you don't have one already.

  1. Thomas Barnardiston
    Thomas Barnardiston 23 February 2023, 09:44 AM

    I would be interested in how AI could be aided to help in aiding decision making outside of the typical 3 options. Having outside information validated and looking back at past projects which encountered similar challenges.