The BIGGEST problem in construction
Sometimes a technology innovation changes the world.
What is the biggest problem in construction? Is it unbalanced contracts, is it unforeseen factors, is it inadequate training, is it too much scope change, or inadequate risk analysis, or too little pre-construction planning, or under-pricing by contractors, or misleading information from clients or inadequate industry capability, the list is long. It is all of these things, and none of these things.
The single biggest problem for all the stakeholders is that we cannot deliver projects to an expected time and cost. Its that simple.
Expectation gap.
The reason that the problem is a problem at all, is because when actual cost and time outcomes differ from expected outcomes one or more of the parties suffers.
All the players in the construction industry have spent and continue to spend huge resources in attempting to align expected outcomes with actual outcomes, that is to minimize the expectation gap. The efforts have come to very little, as this article from John Holland CEO Joe Barr makes very clear. More rigorously the graph below (Flybjerg et.al) shows that statistical probability that a major infrastructure project will be on time and on budget is 2.8%.
Artificial Intelligence Technology
A solution to this problem that makes even a minor statistical improvement is worth billions to project stakeholders, public and private.
Octant AI is the vanguard of the AI technology future of cost and time control for projects and it is working commercially right now. Developed from a collaboration involving Australian innovators and leading academics at University of Oxford, Octant AI leverages over a decade of rigorous research, adapting machine learning to solve the perennial problem of poor cost and time performance.
This is not thought leadership, it is a current platform backed by performance. The graph below is an example of the real-world commercial performance of the Octant AI platform. The data is taken from 96 infrastructure projects that were in progress and completed between January and October 2019. The vertical axis shows that total value of the 96 projects. The horizontal axis shows percent complete. The graph shows the alignment of actual outcomes (black dotted line) compared to the machine learning driven forecasts (blue line) and traditional good practice (orange line). The final outcomes were not known by us (Octant AI) or our customer until the projects were completed over the period.
Octant AI is much better at predicting final cost than the traditional method of experienced professionals supported by software platforms. This is particularly so in the first 2/3 of the project duration (the time when corrective action can be effective). Better knowledge of the future means better decisions, means better outcomes.
A Challenge
I say to the construction industry. If you are serious about resolving the problem of time and cost overrun, then open your mind to a new way of doing things. We just keep doing the same things we have always done, investing more resources and getting the same poor outcomes. Maybe till now that's the best we could do, but now technology can make a fundamental difference. Look to other industries such as Medtech and Fintech. I am not proposing anything that is not already impacting other industries.
Construction stakeholders already have the seeds of their own improvement in the data that has collected over many years. To use it, the preconceptions of how to control time and cost must be challenged. Sure, keep working on improving the many factors that influence the outcomes, but let’s address the biggest problem right now. Close the expectation gap.
Reach out to us, or to other AI powered project predictive platforms (if you can find some, let us know). It is not that complicated or even costly, but it does require active involvement, and an open mind to a new paradigm that actually works!
The best way to prove it, is to try it. There is our challenge to the industry.
David, thank you for sharing 👍
Nailed it Rafal. COVID 19 black swan.
AI won't help. Project planning is generally woefully inadequate and primitive. Reliable methods and tools have been around for decades. Construction contractors do not understand or use them.
Really good read David! Someone might have the impression that you are shearing some profound insight which may give construction companies meaningful advantage over the competitors. The sad true is, construction companies management is in most cases illiterate about IT technology. To implement solutions you mentioned you first need to have some contract management tools or ERP and corresponding habits to feed system with reliable data. This itself is a huge challenge. Adding AI to this is far beyond most managers imagination. To prove your point I would add, that more companies would use such tools, more stable and foreseeable the construction market would become.