How AI Will Advance This Year
Three advancements -- in generative AI, adaptive intelligence, and cognitive computing/NLP -- are just around the corner. We illustrate their benefits and use cases using the construction industry as just one example.
- By Senthil Kumar
- March 27, 2023
The last few years have been challenging for the enterprise. The COVID-19 pandemic caused massive disruptions whose aftereffects continue to ripple throughout the economy today. The supply chain is still unsettled, labor markets are tight, and inflation woes continue. Add in the geopolitical instability caused by the events in Ukraine, dissension in Asia, and extreme weather events caused by climate change, and you have a challenging, unpredictable environment in which to do business.
In dark times through history, human ingenuity has always triumphed over challenges. Some of the best breakthroughs in science, technology, and business innovation have emerged out of crisis. Lean manufacturing methodologies, for instance, were developed by Toyota to deal with the extreme parts shortages the company experienced immediately after Japan’s defeat in World War II. Likewise, during the first year of the pandemic, technologies to support remote work such as video conferencing rapidly advanced their capabilities and reach.
Today, technological advances are accelerating digital transformation to handle current challenges. Markets right now are volatile and changing rapidly. Certain businesses lack skilled labor or the agility to adapt rapidly to meet market needs. Machine intelligence and intelligent machines will bridge this gap, aiding human workers to make them more efficient and effective.
There is perhaps no industry better positioned to benefit from advancements in AI/ML than construction. Historically a laggard when it comes to adopting technology, construction has yet to capture many of the advantages that automation can bring. Additionally, there are huge gaps in information sharing between key stakeholders, and significant inefficiencies with logistics, labor, project planning, and the supply chain -- all of which can be addressed thanks to AI and ML advancements.
Three advancements await us in AI this year. Our company focuses on construction and so our examples are drawn from that industry, but it is just one example of an industry that will benefit.
Generative AI for Business
Generative AI has garnered industry and public interest for the potential it offers. Although adoption of this technology in business environments may be limited, it will soon see rapid growth.
In times past, we wondered whether we could create machines that can think and produce intelligence mimicking human intellect. Generative AI is on that traversal path.
We have already seen AI generate credible programming code, realistic images, and human-like prose besides the ability to answer countless questions. In construction, generative AI will be employed to generate building models, make changes to design, work in tandem with building information models to perform spatial and geometric simulations, generate scenarios to test digital models, and ensure adherence to regulations and environmental guidelines.
Another persistent challenge general contractors face is working in an environment where unforeseen challenges are the norm. A crew doesn’t show up, the weather turns bad, supply chain disruptions occur, an accident happens with a piece of heavy equipment, and so on. There’s always something waiting in the wings, ready to toss a wrench into the gears and throw the carefully planned schedule into chaos. Because these crises don’t just cause discrete problems, their effects have a percolating effect throughout the project. Predicting how they will change the schedule near term and down the line is fraught with ambiguities.
Enter adaptive AI, which can gather all the relevant information from internal and external sources, factor in real-time and real-world signals, perform a number of simulations to foresee the impact of delays and shortages, and predict and adapt the model’s interpretation to fit the new circumstances. It can go beyond simply identifying potential problems. Adaptive AI can suggest ways to mitigate the risks, offer situational intelligence, and improve the odds of delivering the construction project on time and under budget.
Adaptive AI works hand-in-hand with human beings to augment their decisions in situational contexts. It’s extremely valuable, because when well implemented and trained, the model can adjust for unanticipated circumstances and provide highly beneficial and contextual solutions.
Another challenge in construction is the lack of organized institutional knowledge because it is disparate and uncorrelated. Additionally, older employees, who have a tremendous amount of expertise in their trade, are retiring, and there are not enough new employees filling the ranks behind them. As experienced employees retire, much of their knowledge is lost.
Natural language processing (NLP) and cognitive computing -- in tandem with pretrained transformers -- enable an intelligent system to process and understand information presented in multiple formats across multiple sources including structured data, unstructured data, and natural human speech. Cognitive computing and NLP can synthesize tribal knowledge, human intent, and historic information that’s scattered throughout documents within the organization, such as operational efficiency enablers, structural notes, domain-specific knowledge, notes on what worked well and what didn’t work in past projects, and the like. AI’s cognitive capabilities can help synthesize all this information, correlate the disparate pieces, understand the human intent, and formulate a relevant solution.
Further, these technologies ease communication with a wide array of other systems, which can help address multiple construction process inefficiencies. Simply finding good information on the current state of affairs on these complex projects is difficult. Using intent-based computations and the systems’ cognitive capabilities along with NLP techniques, these technologies can enable a construction manager to converse with the system in a natural flow. For example, a user could ask the system, “What’s the status of this project?” and the AI-enabled app not only understands the request and the context but knows from which sources to pull data and assemble it into a form that addresses the user’s question.
These technologies can correlate information from a multitude of sources (such as building design, macro- and microeconomic indicators, adverse climatic indicators, and civic and regulatory compliance) to help compare plans and even real-world objects to ensure objectives are met. It’s an enormous time-saver that also ensures a higher degree of accuracy.
Finally, AI can explain the causality of certain situations and suggest remedial measures users need to take. AI can generate appropriate recommendations and help enforce them.
Construction organizations also need to find errors and mistakes as soon as possible. Errors that are not discovered early can be expensive to address at later stages. An intelligent system can understand the intent of the architect’s design and help address structural and human implementation inefficiencies and errors early when they’re easier and cheaper to fix.
Let’s say that a crew needs to install a window panel and the design of the pane and the drawings are not in sync. Is it a design flaw? A worker error? A material issue? AI can analyze the relevant information -- the RFI, inspection documents, drawings, and other input -- to identify issues and notify the appropriate people and make recommendations to mitigate the issue.
A Final Word
We are witnessing a seminal moment in the evolution and adoption of AI. With the advancements in AI, be it generative, predictive, or probabilistic, the construction industry -- and your industry -- are poised to make giant strides.
Senthil Kumar is the CTO and head of AI at Slate Technologies where he heads a global technology organization focused on delivering the most modern software and technology approaches to leverage data across building production. You can reach the author via email or LinkedIn.