Recently, Bridgit’s COO and Co-founder, Lauren Lake, had the opportunity to talk with René Morkos, CEO and founder of ALICE Technologies, about construction digitization and the industry’s future.
Table of Contents
If you’re unfamiliar with ALICE Technologies or Morkos, here’s a quick overview:
- René Morkos is the Founder and Chief Executive Officer (CEO) of ALICE Technologies Inc. and a Professor at Stanford University.
- Morkos invented the world’s first Generative Construction Simulator, which ingests BIM models and automatically generates millions of valid simulations to build a project.
- Through ALICE, Morkos has already helped contractors reduce project durations by an average of 17% and labor and equipment costs by an average of 13%
- As a professor of Managing Fabrication and Construction at Stanford, Morkos is on the front lines teaching a new breed of construction production management. Using artificial intelligence, machine learning, and other theories, Morkos has changed how the construction production management problem is formulated and solved.
Morkos is one of the leading minds in artificial intelligence for the construction industry and has plenty of insight into the future of construction digitization and contractor data management.
Our conversation with René Morkos
Lauren Lake: We always hear that construction is lagging behind other industries. You’ve probably seen that the very famous McKinsey graph shows that construction is one of the laggards in technology adoption. We’ve always felt that this is a bit of an unfair comparison to make. Do you think this paints the industry correctly? What are your thoughts there?
Rene Morkos: That McKinsey graph states that the dollar value produced per hour of work in construction has stayed the same, whereas, in other fields, it’s gone up. So, other fields have become arguably more efficient. So, that graph is correct. And lots of researchers have looked at it. It’s been validated by third, fourth, and fifth parties, and so on. The question, though, is why? And the consensus is, “Oh, construction folks are not innovative.” That’s the part where… Forgive me, but that’s bullshit, right? And here’s why.
Construction is the second least digitized industry in the world. The only industry less digitized than us is agriculture. As far as I can tell, and this is really from a decade of research, construction just tends to work on bigger, more difficult problems than other fields.
Initially, it’s a broad statement, but being someone who has spent the last 10-13 years of my life working very specifically on construction operations problems, I can assure you that a $350 million gas refinery, like I worked on where you’re burning through $1.6 million a day, represents the biggest, most complex problems that our species has to solve. We effectively build custom-made factories for unique products, sometimes hundreds of miles away from civilization. And that’s not something the other fields tend to do.
The other reason is that to digitize construction, we have to wait for input, which is the design piece, to be digitized. That happened with the advent of BIM, but the issue is that buildings are bigger and more complex than software engines. In my mind, it’s only in 2015 that the processes started to become fast enough that they could actually crunch buildings in 3D, and you can put them in a cloud and then rotate them and visualize them.
So, I don’t think we could really digitize construction until 2015 and later. Interestingly enough, when you look at what I call the construction Renaissance, which is like 2017, I think $140 billion of VC funding poured into construction.
Suddenly, digitization became a sexy topic. When you look at construction as a field, research shows that construction rapidly adopts technologies within a certain sector, such as masonry, electricity, etc. However, technologies that span across disciplines have taken us some time, but my really strong opinion is that it’s not because we’re not innovative; it’s just that we’ve had to wait for the machines to catch up. If you look at the history of digitalization, the first thing that got digitized was finance. How hard is that? We’ve got a bunch of accounts and numbers moving between accounts.
One of the things that I get excited about is if you look at the percentage of GDP represented by the finance industry. In the ’80s, finance represented 10-12% of GDP. Today, finance represents 22%. The fact that it represents 22% means that the salaries are higher, and people are better paid. What I’m hoping, or what I think will happen to construction, is as we start to digitize construction and introduce more and more efficiency, we will tend to have better jobs that are higher paid for fewer work hours.
Speaking of digitizing construction, ENR #97 Power Construction has digitized its entire workforce planning process with Bridgit Bench, which is helping it use its workforce data to its advantage across the entire project lifecycle. Read about their experience with Bridgit Bench.
Lauren Lake: I want to hear more about ALICE Technologies. You take all of these parameters from different workflows, look at available crews and resources, and create the best-case scenario in terms of all of those data points coming together. From what you’ve seen, are any key data points contributing to a project’s success or failure?
Rene Morkos: So what we did was separate planning from simulation. What’s kind of interesting about ALICE is that the plan is the rules that govern your project, whereas the simulation is the solution to those rules. What we’ve realized is, hey, let the people do what they’re good at. Let them use their gut sense, let them tell me how many tasks, how many crews I need, how many crews are available, what’s the radius of the crane, what’s the calendars, those are rules.
Let the computer crunch the other half of the puzzle, which is a simulation. If I have 10 crews available on 20 tasks, let the computer figure out which order I’m going to do them in. That’s kind of how our system works.
To answer your question from our perspective, what dictates a successful project is having a good idea of how I want to build it in the first place. And that seems obvious, but you’d be surprised. For example, we were brought into a $300 million bridge at one point, and I remember one of my colleagues saying, “Hey, they need to start construction in 30 days, but they don’t seem to have an idea of how they want to do it.” This is a problem. And so, to answer your question, I could give you a canned response, which is, yeah, you need to have the list of tasks, the durations, or production rates.
But maybe from a deeper perspective, you need to know how you want to build it. There’s a misconception around ALICE. People think, “Oh, it’s AI.” So I’m going to open it and say, “Hey, how do I build a hospital?”
It doesn’t work that way. It asks, for example, if you want to build a hospital, how big is it? What does it look like? Give me a 3d model. What are the tasks required to build a column? Let’s apply that to all the columns. We stopped calling it AI, artificial intelligence. We started referring to it as IA, or intelligence augmentation because you’ll repeatedly see that you need humans to set up what needs to be crunched and then interpret the results.
If you don’t have any construction experience, ALICE, as a solution, is not useful for you. It’s a multiplier of your underlying ability. If your ability is zero, it’s going to multiply by zero. One of the nice things about ALICE is that it forces project teams to be clear about what we know and what we’re assuming because we’re simulating that construction project for you.
With ALICE, you’ve got to be clear: “Hey, here’s how I want you to build it. Here’s the way I want you to build.” And so that’s kind of what we’re seeing. It’s having that knowledge, that underlying knowledge, before you enter into the project.
Lauren Lake: The construction industry is becoming increasingly competitive. How do your customers leverage their data to gain an edge?
Rene Morkos: We were just involved in a $300 million bid with two general contractors. They had a buffer and a contingency plan of X million dollars. We were working with one of the two GCs in the joint venture, and by running the project through ALICE, they realized that they could half that contingency. What’s interesting about it is that I kept thinking that the objective of ALICE is to find the best solution. I.e., to tell you, “Hey, the best way to build your product is to build it with two cranes, no overtime, regular grind concrete,” and so on.
We started to realize that humans are good at gut sense. So the big deal with ALICE is that you can say, “Add a crane, re-simulate. Try over time, re-simulate.” Humans would take that collective group of 20 solutions and start to understand how their systems behave. For example, a GC thought that the two-crane option would be better, but they’ve run it many times and can see that the three-crane option is what they’re looking for, or vice versa.
To make a long story short, they took that half contingency and won the bid.
Lauren Lake: That’s an impressive story. Is that using data they had been collecting for many years? Or what was the timeline? Do you know how long they had been building up that database?
Rene Morkos: I know they were using databases of production rates and durations that they accumulated. My advisor at Stanford was a guy called Martin Fisher. And so, about three years ago, Martin’s big message to the industry was, “Hey, we don’t have a worked-out data model.” In fancy academic terms, it is a conceptual process model. So, what is a conceptual model? What are the pieces of the puzzle? And so here’s a question I would like to ask my students: Is the production rate an attribute of the task? Is it a property of the task that you’re working on, or is it a property of the crew?
Or is it the property of the element being worked on? The truth is it’s all three together. Something interesting is that if you look at the design world, you’ve got several data models. Interestingly, we don’t have that worked out in construction, which is kind of surprising.
In design, there are probably about 3000 people around the world working on these design standards. We don’t have that in construction. One of the things that we’ve done with ALICE is actually build the first fully worked-out data model for construction simulation.
In the design, you’ve got elements like columns, slabs, pipes, and whatever they are. Each of those elements has some construction information related to them. Up until now, the construction information that pertains to these elements has been across three or four data silos. On the other hand, with ALICE, we have these things called recipes. Recipes ask, “What do I need to build this given element?” Each recipe then contains all the information required to build that element. As a result, I think that will enable people designing operations management systems to far better trade, exchange, store, and learn from data for the construction project.
From the GC perspective, the game is becoming about how you take these technologies and integrate them. ALICE is a part of the solution, but contractors need to look at other solutions and then aggregate those together. They need to really think about the interfaces between those solutions, what data is being generated, where it comes from, and where it goes when you’re done. That’s the kind of stuff that you can start doing today.
Interested in learning how general contractors are integrating new technologies? Read about how The Lemoine Company has taken workforce planning to the next level and plans to integrate Bridgit Bench across its existing tech stack. To learn more about our integration options, check out our integrations page.
Lauren Lake: How do you see the world of data and analytics in the construction industry evolving over the next five years?
Rene Morkos: I think we are living through the most exciting period in construction history over the last two millennia, and it started happening in 2017-18. I believe that machines and technology have reached a point where they can represent the things we need to represent and crunch the problems we need to solve. If you look at companies like Bridgit, ALICE, Versatile Natures, Doxil, and Canvas, more and more of the construction universe is digitizing. As more of it gets digitized, more digitalization can be built on top of it. And I think that’s a great thing, a great day for everybody.
I believe that there’s a new ecosystem that is coming at us in this field. That new ecosystem will be decided by 30-50 companies today that know what they’re doing. Those 30-50 companies are general contractors, designers, software companies, etc. I believe the fundamental value proposition of general contractors is starting to change slightly. Because I think that the GC’s future success will be based on the ability of the general contractor to identify, evaluate, and assimilate new technology. One of the questions I’m often asked is, “Well, why does a company like ALICE or Bridgit exist?” The answer is that our companies represent an outsourced R&D department of these larger GCs. That’s what we are.
The game’s name has started to become: how do you leverage these outsourced R&D departments? And the way to do that is you’ve got to be clear about, okay, well, if I take these two or three solutions and put them together, how will it fit? How will they interact and play nice with each other? how will they play nice with my existing ERP system?
Another thing that I think will start is understanding how to compare group options. If we can generate 30 different building methods, how do I compare these 30 versus those 30? Those are the two big ideas. One is data analytics – understanding how all these complex variables interact and being able to visualize it in a way that makes you understand which are the critical variables. The other is, how do you start comparing groups of options to groups of options versus comparing one to another? Those are two areas in which I see construction data and analytics evolving.
