The prevalence of project failures is a well-known fact in the construction industry – failure typically with regards to cost and schedule overruns, but also with respect to failing to achieve the initial project scope and goals. The default root cause of this is traditionally seen as poor execution and less-than-optimal productivity – generally a reasonable assessment, but only if execution is being measured against a fair and realistic plan or forecast. What if it’s being measured against a plan or forecast that is unrealistic or fundamentally flawed? In that case, execution may be as good as it’s ever going to get, and instead it is the expectations and assumptions upon which the plan is built that are off.

This question challenges the traditional relationship between project planning and project execution. It all boils down to knowledge and continuity of information flow between the two efforts. Historically, these two exercises have been separate and supported through the use of two very different disconnected toolsets and, hence, disconnected data.

With the increasing digitization of project data and the emerging application of artificial intelligence, the challenge of alignment between a realistic forecast and achievable execution is starting to be overcome.

How Does Artificial Intelligence Help?

So, what if contractors could improve their ability to establish more accurate forecasts earlier on? What if they were able to easily look back and plan more accurately based on actual historical performance and trends?

To accomplish this, the interaction between human and computer needs to change. Instead of the human simply feeding data into the computer (e.g., building a plan), the computer needs to assist the human and actually make suggestions. By changing the direction of communication from being solely human-to-computer to more of a bi-directional flow, the human can then benefit from the knowledge, history, performance and trends stored on the computer.

This is exactly where the science of AI is changing the construction project landscape, and it’s only going to accelerate. By capturing planning knowledge such as schedules and cost forecasts, as well as performance execution information such as as-built plans, organizations are better positioned to develop more realistic forecasts.

Artificial intelligence algorithms enable contractors to mine historical information and offer informed suggestions based on history. The magic to such an approach is the understanding of context. When the computer makes its suggestions or offers guidance, it needs to understand context pertaining to the type of project, geographical location, scope of work, quantities involved and so on. The more points of reference, the stronger (or smarter) the computer inference can be when it mines data from its knowledge library. This eliminates the need for highly uniform data sets even to the extent where common terminology is not required for the computer to make an intelligent match or suggestion.

The days of non-guidance-based project management software are limited. Next-generation guidance-based tools for both planning and field execution are quickly becoming the norm.

The Knowledge Digitization Trend

The project management industry has evolved through the development of numerous tools with a very specific purpose (e.g., scheduling, risk analysis, earned value, statusing, etc.), each coming from a different vendor. As a result, data flow between these point solutions has always been a challenge.

With the advent of cloud computing, the industry is trending away from discrete point solutions and towards a “single source of truth” type of environment. This, in turn, is enabling much easier mining of data through the use of AI. With increased amounts of project data being digitized and more centrally stored, AI algorithms can focus on intelligent mining of knowledge rather than unnecessary effort in consolidating disparate data sources. Historical data is quite simply more accessible today than it has been in the past.

With the abundance of digitized project knowledge comes the need for knowledge cleansing. Normalizing knowledge to eliminate bias and removing outliers so as to avoid erroneous suggestions are essential parts of establishing a useful knowledge repository.

The Owner/Contractor Firewall Is Crumbling

Another emerging trend is that of a higher degree of transparency and collaboration between owners and contractors. This has been partly driven by the introduction of integrated project delivery (IPD) philosophies. The thinking here is geared toward placing a greater emphasis on the value to the owner in terms of everyone focusing on the overall asset being built rather than each discipline solely focused on executing their own respective scope of work. In addition, collaboration incentives further reduce wastage, especially in the design phase of the project.

There is a very important human element here that complements the AI aspect described earlier. Encouraging collaboration by establishing consensus of opinion between owner and contractor or allowing the contractor(s) to have more of a say when the forecasts are being developed is resulting in more buy-in, which is a huge step towards overcoming the productivity challenge.

Does AI Mean the Computer Is in Control?

Leveraging the computer’s ability to offer intelligent suggestions does not mean contractors are losing control. No matter whether the construction employee is a planner, estimator, superintendent or site engineer, their expertise is still paramount to AI success. The computer and its AI capability really only augment, and certainly don’t replace, human expertise.

Taking this to the extreme, the computer can itself benefit from the human through machine learning. AI algorithms can automatically adjust and recalibrate based on human expertise. This is where the real power of AI and human intelligence comes together. If the computer can do the work that humans find time-consuming, such as mining through large data sets, and humans can then train the computer to make more informed decisions or suggestions, things gravitate towards a continuously improving circle of knowledge and intelligence.

Conclusion

Project expertise and knowledge have been present for decades. The challenge, though, has been to capture, retain and reuse that knowledge to the advantage of other projects. With AI becoming more mainstream, the ability for domain expertise and lessons learned during execution to be captured in a way that can subsequently be interrogated by a computer is now a reality. This reality means that projects no longer have to be planned in a knowledge silo devoid of buy-in and consensus. With the calibration of AI guidance and the validation from domain experts, the resultant plan against which the project is executed is, without doubt, a more achievable goal.

Want to see what your future projects could look like? Contact InEight to request a free demo and learn how AI-powered planning solutions such as InEight Basis can help you transform your planning and scheduling processes so you can plan smarter, not harder.

Request a Demo of InEight AI-powered Project Planning Solutions