Why AI and machine learning belong in the construction and engineering industry

In our latest Trailblazers interview, we speak with Roz Buick, Oracle’s senior vice president product, strategy, and development, Oracle Construction and Engineering.

Buick talks about her career background, shares her recommendations for cultivating innovation in construction, where AI and machine learning fit into the industry, and the value of streaming data from one project to another.

Dr. Burcin Kaplanoglu, vice president, Oracle Industries Innovation Lab, led the discussion.

What's your current role and how has your career evolved since you started working in the industry?

I'm senior vice president of strategy in the Oracle Construction and Engineering Global Business Unit. We've got a global team who's working hard to innovate and drive a next generation digital platform where we bring a lot of our core applications together to serve owners and delivery teams.

I was originally trained as an academic research scientist. So early AI and neural networks and various algorithms were integrated across a problem, whether it was watershed management, environmental, or land use planning.

I survived the first round of AI, I like to say, back in that phase. And then I switched to more of a commercial company focus and had more than 20 years working across hardware, software, and services to integrate and innovate how people do their work and focusing in on automating machines, particularly in agriculture and construction. What was fascinating about my journey is that now the AI field has come back, so it's an interesting story.

What's your view of the state of innovation in the construction and engineering industry?

Innovation is an all-time high, and there's more money than ever flowing into the industry for startups and semi-established businesses. It's the best time to be in this sector.

There’s also a number of key established players in the industry, like Oracle, which are seeing significant market adoption at this point in the industry. This signals that the trend toward going digital is becoming more and more urgent in many of our customers and together, all of that is fueling the industry to the next level.

Since it's a great time for adoption, what are the biggest challenges to innovation?

The biggest challenges are cultural and change management. I think companies struggle to know what going digital means. It can be quite overwhelming. There needs to be a balance between that decentralized empowerment at the project level and a centralized standardization of processes that are core competitive areas for that company.

Senior executives of every company in construction and engineering need to understand how to build a data strategy for themselves. Like any good Lean Six Sigma transformation in a company, you need buy-in from the top down.

Pulling data and metrics around your core business processes should already be defined in your core strategy. Selecting your digital partners stems from that, but it really begins with human beings and cultural shift in a lot of companies. The acceleration of what we're doing is less about individual technologies, and as much as integration of technologies to solve bigger problems.

See what’s possible with the Oracle Smart Construction Platform

The fact that you have processing in the cloud, the bandwidth, the mobile, the sensors, the computer vision, the AI, and automation technologies that have reached levels of sophistication now and coming together in various ways that is driving a massive shift. That is the opportunity for everybody.

How can organizations foster a culture of innovation?

I really feel my experience with data and information has been a central anchoring one for me. It's important as we go forward, particularly around AI, that as you're building every project, you need high-quality data.

I learned this by being a research scientist over many years. You couldn't publish data unless you had drilled into how you design the hypothesis, how you design the experiment, the methodology of the experiment, what were your study endpoints or outputs, the measurements that you are generating, and why and how it all relates. You couldn't get through that gauntlet without peer review of your publication to generate data into a community of scientists and experts around the world.

I feel data quality is often overlooked in this frenzy to get to AI and machine learning. Data is the next oil, but not if it's bad quality, it hasn't got good continuity, it's not clean, or it's not secure.

These elements of data integrity tend to get overlooked. It’s important to the future of AI and to be able to deliver the promise of AI—the data is kind of like the bedrock, or the saying, "garbage in, garbage out."

If you're banking on bad or missing data in your overall modeling, no matter what algorithm you apply, it won’t be good.  When you're applying data to an algorithm to create a model, and that model is then becoming your guiding light to predict intelligence or insights, you must have a strong bedrock of data quality or integrity, otherwise it's a risky place to build your home.

If you're building a home in an earthquake-prone country like Japan or New Zealand, there's no way you'd build it without being earthquake-proof. Quality & security are becoming this earthquake-proof data standard now for data in the cloud.

Every organization adopting a data strategy needs to think through their differentiating best practices to compete in the market. How they create the metrics and data around those competitive processes is foundational to moving forward and continuously improving.

Pick a digital set of tools that will allow you to build your standardized practices in a decentralized versus centralized balance of processes that then lets you access that data well through a common data environment. A CDE that allows people to access any dimensional data, whether that's flat files, relational databases, or 3D, 4D, 5D models and so on.

You need to be able to find a partner that can build your digital platform backbone that takes the common data environment architecture from the very beginning into account. That lets you build these AI machine learning models off that common core data environment.

Assuming the data quality is there, you're on a continuous improvement journey for everybody to build these models, improve, and iterate. As an executive team, not only do you need to sponsor this from the top, but you also need to adopt the principles of science and data science, which is to continually debate and challenge each other about what is right and what is not right. Are the models or the data integrity where they should be because that is a journey in of itself.

What emerging technologies do you see presenting the best opportunities for the industry?

AI and machine learning is a key opportunity for us. But we need to be careful how much we oversell or set expectations. We're not replacing human beings; it's about ensuring that humans can do their job better so they can do more interesting and challenging intellectual problems beyond that.

More and more companies are going to struggle to find talent to work in construction. So having remote jobsites, computer vision sensing, IoT sensors around resources and bringing these kinds of data into your central platform to manage that data has got to be a focus going forward.

Pick a digital platform to handle that data in a common data environment, and then you can mine that intelligence with an analytics and AI machine learning tool such as the Oracle Construction Intelligence Cloud Service that we're building out across all our applications on the platform. Augmented reality, visualization and computer vision combined with sensors is a huge space that will expand the kinds of data being used. Because of this combination of technologies, we're seeing a huge acceleration, to say nothing of on-the-jobsite robots like Spot the dog that we all know from the Oracle Industries Innovation Lab work we have been doing.

I struggle to see enabling robots or automatons removing all work for humans. As in round one with AI, the expectations were set too high in replacing human beings.

That big thinking about AI generally is a false goal. It's trendy and great for media articles, but the reality is we still need to be focused on AI to help workflows or business outcomes and not reinvent humans.

In our industry we can see a ton of opportunity for AI and machine learning. But again, the data foundation is core to empowering all of that - how you store data, how you ensure integrity and security of the data.

Regarding AI and machine learning, what would be some use cases where you see the impact?

There are many. Within scheduling, predicting schedules that are looking riskier for a normal practitioner who wouldn't even think of spotting these problems.

AI algorithms and machine learning models can pick up these trends based on past project data, in the current project, and into future projects so you can be smarter earlier to spot these issues. Over-running costs can be another area for predictive intelligence.

The idea of continually improving can be a model in and of itself, and you get more learning over time as to how those behaviors and good schedules evolve. Building AI models around that is another area with opportunity.

Combining that with your core process like scheduling and project management to keep the entire organization on track and be smarter earlier about potential issues. I think they are some of the big ones.

There's a lot of data, but about 90 percent of it gets left behind from one project to the next. The biggest gain in value is how do we take that data from every project or take from one project to the next to build a model. Taking that consistent level of continuous stream data across model projects and feeding that into these machine learning models that then help you run projects as a team a whole lot better.

You want to connect, you want to empower, so collect and collaborate through a project collaboration application. But empower through mining the common data environment and using AI machine learning through tools like Construction Intelligence Cloud Service.

Then synchronize more parties across stakeholders so they know the changing variables on a project. If you're using one digital silo tool, you don't get to see that.

If you're a scheduler, you don't know what the payment activity is over in Oracle Textura Payment Management and where the progress on that is with a vendor. You could be connecting a lot more of that information between constituents, then giving them insight and intelligence earlier. Be smarter earlier is sort of a catch phrase, but there's almost a limitless set of opportunities on how you apply AI.

We're on a journey to get better at this as the whole industry works together. If you're a client thinking about data strategy, data integrity, and quality data security alongside us, we can make this industry move a lot faster. It's a journey together.

Check out more LinkedIn Live sessions.

Oracle Construction and Engineering, the global leader in construction management software and project portfolio management solutions, helps you connect your teams, processes, and data across the project and asset lifecycle. Drive efficiency and control in project delivery with proven solutions for project controls, construction scheduling, portfolio management, BIM/CDE, construction payment management, and more.

Related posts

• LinkedIn Live from the Oracle Industries Innovation Lab