Will AI replace Project Controls professionals?

“One point that (is) made is about the uniqueness of humans when it comes to evaluations. It was called judgment here; in my jargon, it’s evaluation — evaluation of outcomes and basically the utility side of the decision function. I don’t see any reason why that should be reserved for humans” Daniel Kahneman 2018

What is project controls?

In a profession which attracts people who thrive on detail and have high levels of system focus it is little wonder that a concise definition is often a matter of debate. For the purposes of this article I rely on a definition penned by Pat Weaver : Project controls are the data gathering, data management and analytical processes used to predict, understand and constructively influence the time and cost outcomes of a project or program; through the communication of information in formats that assist effective management and decision making.

The whole purpose in making any decision in any circumstance, is to achieve a hoped for future benefit arising from the implementation of the decision. To the extent that the outcome of implementing the decision is predictable, the risk of failing to achieve the benefit may be reduced. So a core value of project controls is to reduce uncertainty through prediction and understanding.

The current state of the art.

In 1995 I proposed to a client that the project we were “rescuing” needed a project controls manager. The response was “A what?”. Over the past few decades we have observed the emergence of Project Controls as a specialty in the pantheon of project management services. But with a focus on data, prediction and analysis and the increasing complexity of major projects, AI will inevitably impact the role of project controls professionals.

Recently I have noted many comments by very experienced Project Controls professionals along the lines of “Machines, can never replace our professional expertise”, “All projects are unique, so it is necessary for experienced professionals to apply their unique judgement” and so on. Naturally professionals who have spent a career developing high levels of valuable expertise are reluctant to consider that their skills can be replicated by a machine (as was Garry Kasparov).  Yet to hold the view that current best practice project controls will remain unchallenged by AI is naïve. We know this because we have already developed an AI platform that is better at accurately predicting forecast final cost and time, than experienced professionals, and which is being deployed commercially.

Forecast Cost to Complete for Portfolio – All Projects January – October 2019.

The graph above is compiled from actual projects and it is obvious that machine forecasts more closely mirror actual final costs than the experienced and proven project professionals involved in this portfolio of projects.

A very common view of the impact of artificial intelligence is that it will replace routine work, but that project based knowledge work which currently requires expert professional judgement is not a candidate for AI impact or that this AI capability is in the far future. In the 2017 Arup report “The Future of Project Management” (Rob Leslie-Carter et al) the timeline identified 2040 as the time when “Smart algorithms proven to be better than expert judgement”. The prediction is accurate, but 20 years late, as the graph above shows.

The future of Project Controls

A new paradigm is emerging, and as with all paradigm change the traditional perception of reality changes. In is recent book “The Technology Trap”, Carl Benedikt Frey distinguishes between “enabling” technology and “replacing” technology. Project Controls has to date been supported by enabling technologies (such as P6, Deltek Cobra, Prism etc) which have also been core technologies in its emergence as a specialty. These enabling technologies allow experts to leverage their knowledge and skill. It is reassuringly tempting to simply see AI as one more enabling technology. However, the risk of failing to recognize a “replacing technology” and adapt accordingly, may be a 2012 “Kodak moment” (look it up).  Project Controls professionals who will thrive in the not so far away future, will be those who understand the implications of this, and embrace “augmentation” based on a partnership with new AI technologies.

This will not be easy, since there are many different sorts of AI and applying the right form in the right circumstance using the right data is a critical component to getting results like those we have achieved. Canny Project Controls professionals will embrace machine learning and develop enough knowledge to guide their clients. They will become informed collaborators with artificial intelligence technology. This means a change in self-perception, recognizing that expert judgement and prediction can be provided by machines more accurately, faster, earlier and cheaper. Augmented Project Controls professionals will be presenting and explaining insights similar to those you see here: https://www.youtube.com/watch?v=ah6MKwjZbMc .

So for the sake of debate, I offer a new definition.

Augmented Project Controls is the combination of traditional data gathering, data management and analytical processes with leveraged artificial intelligence generated prediction, planning and adverse event protection, to constructively influence the time and cost outcomes of a project, program or portfolio through effective management and decision making.