Joe Fuller: Designing anything—from buildings to machine parts—is all about accounting for stresses and loads and creating resilient structures. Autodesk, a pioneer in computer-aided design, is closing the gap in sectors like construction and manufacturing between designing and building products, and the potential for disruption is obvious.

Welcome to the Managing the Future of Work podcast. I’m Harvard Business School professor and visiting fellow at the American Enterprise Institute, Joe Fuller. Today, I’m speaking with Autodesk president and CEO, Andrew Anagnost. Andrew is a PhD aerospace engineer who was promoted to the top spot at the company in 2017 after several decades in design and engineering roles. As it pushes the envelope of design-and-build technology, Autodesk finds itself in the middle of substantial shifts across a range of complex industries. Andrew and his colleagues have made a point of factoring workers and communities into the design parameters of their products. That’s a tough task. A techno-optimist, Anagnost argues that automation can increase opportunities for meaningful work while it increases the productivity of our society.

Andrew, welcome to Harvard Business School and to the Managing the Future of Work podcast.

Andrew Anagnost: Thank you, Joe. It’s great to be here.

Fuller: Well, Autodesk has had a front-row seat at the evolution of technology as it affects not only companies’ work but entire value systems of work, different entire industries. How has that changed over time, and how do you see technology affecting work for your customers?

Anagnost: Autodesk has a unique position in the world of making things. We’re used at the front end of just about everything that’s made: Buildings, bridges, airplanes, cars, machines …

Fuller: … apparel.

Anagnost: You probably know we started off as a 2D drafting company, and that’s what we’re most known for. And that, in itself, was a pretty robust revolution. It changed people’s jobs. It eliminated entire categories of work. It created new categories of work. There’s actually more people designing things after the automation of drafting than there were before. But today, we’re much more about being a modeling company—although we still do drafting software—modeling 3D representations of designs, which is another big disruption in how people do things in terms of visualizing their designs, capturing their intent, and actually communicating into the supply chain. But what’s happening now is probably even bigger than what’s happened in the past, and it’s what we like to call the convergence of design and making. And by “making,” I mean specifically how you build something, how you manufacture it, or how you produce it if it’s a film or a game. What we’re moving to is a world where these 3D models are going to be living and breathing things that facilitate very broad and very deep collaboration between disciplines and very high levels of transparency into what’s happening on the make side vs. what’s happening on the design side.

Fuller: You are jumping over all sorts of different historical lines of demarcation between companies, between industries, who does what, who owns what. How do you keep control of that? And what’s enabling that to happen now uniquely as opposed to why didn’t it happen 15 years ago?

Anagnost: There’s kind of a spectrum of process integration and process digitization or process alignment or supply-chain fragmentation. At the front of the spectrum, believe it or not, is the film and game industry. They are actually highly integrated, very tightly integrated delivery changes, highly digital processes. And then you move into manufacturing, which most people would consider the bellwether, where you are deeply integrated, the processes are pretty mature, supply chains are well established. Large, what we call OEMs—original equipment manufacturers—kind of dominate their supply chains and kind of drive control over lots of various things. And then you get into the architecture, engineering, construction [AEC] world—highly fragmented, very early in its digitization processes, not mature in terms of aligned processes. So this spectrum is like moving across all these industries. And what you’re seeing is some of the manufacturing-type mentality is now making its way into the AEC world. There’s a couple of things that are rising up and making them reevaluate how they do things. 3D printing is going to be a ubiquitous and robust technology. That is forcing companies across the whole spectrum of built things to reevaluate how they actually build something. The other thing is the rise of multipurpose, intelligent, configurable robots. They can actually do diverse tasks and can do them in ways that mimic human behavior or mimic kind of the processes that humans have to go through. And then the last piece is the rise of mobility and tablet devices in just about everything our customers do—across the front-end processes all the way down to the construction site, the manufacturing floor. These devices have brought a level of technological competency to places where they simply did not exist before. Now, as a software vendor, we’re able to do things computationally within the purview of the applications we deliver that we’ve never been able to do before simply because the cost is going down. Now you layer on top of that one more fundamental technology: Machine learning is allowing us to automate things that before we really wouldn’t have been able to automate in any simple way, because we can capture rules, we can capture processes, and actually use that captured knowledge to automate things. When you combine all this together, this is what’s driving the change for our customers. Technologically, we’re delivering new types of automations, but their end markets are being fundamentally disrupted.

Fuller: Andrew, what are you hearing from your customers about the ability of their workforce to keep up with these changes? Because in all of our research here at the Managing the Future of Work Project, what we keep hearing from employers is, “I’m deeply worried that my current workforce is out of phase with the requirements of work as they’re unfolding. I’m deeply worried about my ability to get the right type of people.”

Anagnost: So the technology and the process change often bubbles down to how the people can implement it. So we encounter all sorts of angst in our customer base about where things are going. Right now, in both of our end markets—construction and manufacturing—people are having trouble finding people to do the work that needs to be done.

Fuller: Worldwide.

Anagnost: They’re super-concerned about the skill sets they’re hiring and about the ability of people to adopt these new types of technologies. They’re also very concerned their most-talented individuals are aging out of their workforce. They’re not getting enough of the next generation. And a bunch of knowledge is starting to age out or walk out. So they’re working with us and internally to try to capture some of this knowledge and codify it before it’s gone.

Fuller: So is your design logic to create technology that the people who are actually available to do the work can handle and be productive using? Or is it to maximize the productivity of the process and offer the best possible solution to your customers and let them wrestle with how to staff it?

Anagnost: Yeah. So it’s both. If we’re going to integrate cloud computing and machine learning into a platform, the platform can’t be static. It has to be extensible by the customer. It has to be able to integrate. Their knowledge has to be able to layer on top of it, and we have to be able to process it. So we’re definitely going to provide things that codify the processes for them, that help them move the process from one stage to the next. And we’re starting to build new types of things into the tools. So the tools are starting to watch the individual—we’re multi-years away from doing all of this stuff to the level that we want—but the tools are starting to watch you as a professional. And what they’re doing is, they’re starting to compare you to people who are further down the technology adoption curve than you, and trying to understand what are the things you have to navigate in terms of new technologies that are inside the products or services to get your skill level up to where it needs to be to meet the growing demands of your employer. We then apply that knowledge and turn it into a learning path for the individual so that they can actually learn as they work. So what we’re trying to do is watch and assist as they work, which is a very different paradigm than what we have today, where they go to a class, they try to learn the new tool, and then when a new thing dumps in, they have to pause and figure out how it works. We’re trying to watch, curate, and provide a continuous learning path for them and then automatically badge them—the whole idea of micro-certifications—so that they can actually get recognition for what they’ve learned. We have to help people adopt a technology more rapidly.

Fuller: So you’re drawing data from your entire embedded base of customers and using that to inform the way you train anyone in the kind of skills hierarchy, irrespective of which customer it is or what job they’ve got.

Anagnost: We have a new technology called Fusion 360, where we are actually collecting data on how customers use the application and trying to codify their learning curves and help them progress more aggressively.

Fuller: Now you come from an engineering background. You’ve been CEO for a couple of years now at Autodesk. How does this all play out within the four walls of Autodesk?

Anagnost: We’ve actually significantly rebuilt the HR practice inside the company. There’s this triangle between CEO, CHRO, and CFO now that kind of helps us understand not only how we’re hiring people but how we’re bringing them along, because a company can no longer afford to just kind of throw away entire chunks of their workforce and then hope that they can hire new workforces in to do these things. They have to do a combination of evolving their existing workforce and bring new workforce talents in. We’re actually reevaluating all the systems and processes we use to train people. We’re doing things in much more of an on-demand-training basis. We’re monitoring where people are in various levels. And, by the way, it’s a lot harder to manage someone’s managerial competency than it is to manage their competency on a particular design tool.

Fuller: I bet.

Anagnost: So we’re incorporating data into our processes with regards to our human capital, and we’re incorporating capabilities that allow us to help people move in bite-sized chunks that they can consume on their own. Of course, we’re always striking this balance between how much do we want to know about our employees so that they feel comfortable that we’re actually trying to learn things that help them benefit vs. what starts to look a little creepy. So employees, just like customers, opt-in on capabilities that allow us to understand how they’re working and help them be more productive. So we have a whole initiative to digitize our workforce that’s similar to the kinds of things we’re asking our customers to do. We’re investing in systems, we’re investing in processes, and new types of people with new views of how you train a workforce.

Fuller: I think that’s going to give you valuable insights into how to be more effective serving your customers. Is Autodesk’s workforces suffering from the same graying phenomena you talked about earlier?

Anagnost: We are actually a very different company than we were three years ago. We’re less exposed to the risk of a workforce that was aging out or moving on. We have a much more spread out demographic. But with that comes challenges. The needs of each stage of these people’s careers are very different. And our customers are going to be experiencing this exact same phenomenon.

Fuller: You’ve also begun to expand the business model of the company in some interesting ways—right down into looking at things like prefab construction. Talk about that and the logic for that and how it’s a reflection of the way you view work being managed in the future.

Anagnost: This whole process is moving to much more of an industrial process. It’s a highly wasteful industry. It’s responsible for 30 percent of what we see in landfills today. The figures are staggering. And at another level, people are going to have to build about 13,000 additional buildings a day to satisfy the population growth that we’re going to see by 2050. We’re going to be a 10-billion-person planet. People have to build buildings, they have to build them quickly, they have to build them efficiently, and they have to build them without destroying the planet. There’s not enough natural resources, there’s not enough landfills, there’s not enough capability to do this right now. There’s not enough money. So what we’re moving to is a process where things are actually built like you would build an airplane. You prefabricate lots of components. You preassemble multiple versions of them in a factory. And what the construction site becomes is the final assembly facility. They’re all fit together by a workforce that is now used to assembling high-precision things, just in an outside open-air factory. That change is coming to the entire built-world industry there. Everything we build is going to be moving to a mix of these bespoke processes, but these highly industrial-like processes. This is the way we’re going to be able to afford the buildings we have to build. This is the way we’re going to reduce the waste that’s associated with the industry. This is the way you’re going to see safety and productivity go up at construction sites. Because right now construction is one of the most dangerous jobs you can have out there. It’s going to look lot more like the factories did 40 or 50 years ago. The industry is very interested in industrializing itself. You go talk to the construction professional, a GC [general contractor], or even a large design-build firm, and they’re going to tell you, “We can’t continue like this. Our margins are low. Every project is super risky. Our ecosystem is fragmented. Everybody’s trying to arbitrage the risk across the project so that they don’t have to be responsible for whatever mistake was made.” They can’t continue the way they are right now.

Fuller: And certainly in some markets, availability of labor under the current definitions of work is getting very, very scarce. Certainly in home building in the United States, that’s an endemic problem—finding skilled tradesmen and even unskilled labor to work the construction sites.

Anagnost: Yeah. We’re investing in it because it’s an opportunity, but also because we want to solve the problems that bring these two things together. We invested in a company called Factory OS in the Bay Area, which is basically doing modular home construction in a factory environment. They build prefabricated modules, ship them to a prepared site, and assemble them on site. And they’re doing it for lower-cost-housing applications. But it’s one of the interesting areas where they’re trying to apply it directly to a pressing social problem that’s out there right now, which is, essentially, the price of housing in California and the ability to drive down that price, drive down the waste of creating it, but also at the same time providing high-quality jobs for people whose work was pretty challenging before. Right now they’ve partnered with the Carpenters Union to bring people into the factory site doing this prefabrication. These people come to a single place to work every day, they build these module units, and then they also work with the GC on site to help with the assembly on site. It’s a very different work paradigm.

Fuller: What do you and your colleagues observe about the receptivity of groups like labor unions to this type of technology? Because certainly the meme is very resistant to change, not accommodating it.

Anagnost: There is a lot of resistance. One of the reasons why we were very interested in Factory OS, in particular, is because they managed to broker a deal with the Carpenters Union to try to drive a new work paradigm. I think we need to do more of that. And I think it’s an important part of how we address this change because, yes, the unions in general are afraid that this is a job destroyer for them and it’s going to limit their power. But I think there are other paths where, if they work cooperatively with the right people, we’re going to be able to provide something similar to what happened with the United Auto Workers and things associated with that as the automobile environment evolved into large monolithic factories. As a matter of fact, I personally think the path of partnership is the better path, because they have resources and access that we can tap into to help people learn new skills. The kind of tech-classic model—where you just kind of roll over people and try to do it that way—it’s not the best approach. We really do need to partner if we’re going to bring this workforce forward in the future. There may be fewer people working per project, but the throughput of projects is going to go up in this highly automated world. And we need to get these people along with us, because we don’t have enough skilled labor to manage the throughput we have today. If we’re going to be having more projects, we need these organizations to embrace this. If they don’t embrace it, we’re going to see more of this difficulty of offshore entities bringing in prefabricated technologies and fulfilling the demand elsewhere. And by the way, that’s happened in some respects with some projects, where prefabrication has been done at a very large scale but the modulars themselves were built offshore. That’s not what we want, and that’s not the direction we should be going.

Fuller: If we look at the AEC industry, and also you mentioned earlier an industry in the filmmaking industry—which are industries that have lots and lots of different suppliers of different size and scale—and coordinating them has been a big challenge. But it can be done, as the film industry proves, because, frankly, when people are skeptical about the gig economy, I ask them if they’d gone to a movie, ever.

Anagnost: Movie is a massive gig economy.

Fuller: How do you see technology? Is it going to reward scale and create the type of outcomes you’re describing in ACE, where we may see more industrialization and less of this kind of balkanization and fragmentation of industries? Or is it going to enable a proliferation of industries where you can assemble the right talent at the right time?

Anagnost: Yeah, so it’s going to be a progression. What you’re going to see in the AEC space in particular, is you’re going to see the rise of design-build firms. When you have a highly industrialized process, and you’re trying to tune these processes on a single 3D model that carries all the way through the process, you want very tight relationships with your supply chain. You’ll actually want to integrate from design all the way through to construction, maybe even down into the trades, though you could see the trades continuing to be fragmented. And at the same time, you’re also going to see the cottage industry rise up as well, because if the cost of producing something can be much more managed, and you can provide small scale means of productions—for anything from housing to smaller office buildings—there’s going to be room for smaller boutique businesses as well that capitalize on the technology as well. But short term, you’re absolutely going to see a consolidation shake out a lot of the things that need to change—and not just inside the industry, but inside the government entities that regulate these industries: permitting processes, codes, all the things that kind of wrap around the industry that allow people to actually … land use. All of that’s going to have to adapt as well.

Fuller: Andrew, you’ve talked a little bit about how technology is going to affect lower-skilled and middle-skilled workers, for example, in the AEC industry. What about the higher-skilled worker? What about that—the architect, the designer, the land-use planner—how are they going to be affected by all this?

Anagnost: The engineer.

Fuller: The engineer.

Anagnost: There’s an old joke in the AEC world that the architect designs the building and then dares the contractor to make it. And those days are going to be over. And architects are going to have a lot more exposure to the impact of their decisions on how something is actually built or how something propagates through the process. We’re going to be able to provide it to them through machine facilitation. We’re going to help the disciplines collaborate better together, and we’re going to shorten the distance between these decisions fairly dramatically. At some point, they will be instantaneous. You’ll be able to see the immediate implications of a design decision. That’s going to pull the architect a lot more.

Fuller: I’m sorry, but throughout the entire system, right down to execution in the field?

Anagnost: Throughout the entire system.

Fuller: Wow!

Anagnost: Because if you have a living, breathing 3D representation inside the computer that you are constantly updating and computing as various disciplines contribute to it, everybody gets to see what’s going on at all points during the process. And if the system underlying that is smart enough to predict what the impact of a particular change is going to be and feed that impact, say, to the construction professional, everybody is getting lots of high-fidelity input on how things are going.

Fuller: So much more active dialogue.

Anagnost: Much more active.

Fuller: Trade-offs and constraints.

Anagnost: I like to say that we’re going to be moving from these highly serial, over-the-wall processes to a design-and-make process that looks a lot more like a multiplayer online game, where people are collaborating instantaneously on a lot of decisions, working in their niche, but ultimately kind of seeing what the other person is doing. The discipline of architecture will likely change, because they have to understand more of the downstream implications. They have to integrate them more. Their owners—who are going to be paying them for these projects—they’re going to expect them to have a high-fidelity understanding of the ultimate cost and buildability of their projects. Designers are going to know more about engineering, they’re going to know more about construction. Architects are going to know more about how the construction processes work downstream. It’s not that they don’t today, but it’s going to be much more integrated into their process and thoughts earlier on in the design.

Fuller: One thing we see in our examination of the future of work is that current practices have led to the creation of a legal-regulatory-policy surround. You’re talking about changes that would obsolete a lot of that surround. How much of a constraint is that going to be, and how can we prevent those externalities from gumming up the works?

Anagnost: The ecosystem is going to slow down the adoption of this technology. One of the reasons I don’t worry about people’s jobs going away overnight is, one, I believe there’s going to be more work and, two, it’s going to take longer than anyone thinks. There’s a contractual environment that wraps around that ecosystem. There is a legal environment. There’s a code environment. There’s layers upon layers from local governments to how these various entities in the process collaborate or don’t collaborate. And it’s all codified, either formally or informally. That’s one of the reasons why we invest in companies like Factory OS that are looking at the policy barriers and the contractual barriers to enabling modular and prefabricated construction.

Fuller: One last question. You’re talking about people with postgraduate degrees, highly skilled engineers, architects, right down to day laborers, let alone master carpenters or pipe fitters or the cabinet fitters. What are the kind of core skills they’re going to need in the type of environment that you see unfolding in AEC, unfolding in manufacturing.

Anagnost: Let me tell it by maybe giving a little bit of a “from-to” context, and then I’ll talk about this specific skills. In manufacturing, I said earlier, we’re probably going to be moving away from this world where there are human-intensive factories, where a lot of assembly activity goes into highly automated factories that have very sophisticated machines that need to be maintained, serviced, programmed, and configured. And more people are going to be able to feed those means of production. So the factory worker of the future is actually a manager of a facility that has highly configurable, highly programmable machines inside it that need people to maintain them, configure them, update them, adapt them, and evaluate their performance over time. All right. So you’re seeing a factory worker moving from a task-specific individual or a machine-specific individual to a machine generalist in some respects. And you could even see a point where someone inside the factory is actually dealing with a specific customer problem with regards to how their thing is being built inside the factory. So human skills—the ability to problem-solve for an individual that might be using that particular facility—is going to be a key skill for the individuals that work inside these facilities. It’s a little bit different inside of AEC, because AEC is moving to a world where it’s going to look a lot more like manufacturing did 40 years ago. So what does that do for your classic construction worker? They’re going to have to have some of the skills that an assembly-line worker in an automotive or aircraft factory had 40 years ago. They’re going to have to learn and acquire these skills, where they understand multiple stations, they understand how processes work, they have to get comfortable with higher-precision processes, they’re going to have to be able to use configuration tools and technology tools that allow them to access information about what needs to be done next, provide input back to individuals who are up front in the design process. And that’s going to change the character of the people that are working in that space. Now, I am one of these people that has ultimate faith in all of these people to adapt and evolve to these things, and I think technology is going to allow them to more effectively operate in this world. I have more concerns about the ecosystems that are surrounding them to support them in absorbing those new technologies and applying them.

Fuller: So often we come to the point where we say that we know what people need to be able to do in the future and the basic capacities they’re going to need to have—competencies they’re going to need to have—to be able to take advantage of the type of productivity increases you’re talking about that are embedded in these technologies. But can we, can the education system, can the regulatory system, can the legal system adapt fast enough to put them on a quick arc to realize those opportunities?

Anagnost: This is a very important conversation that I think we all have to be catalyzing, because the education system’s behind. Look at the current policy debate right now. It’s very interesting how backward-looking at is. Okay, we have a large discussion out there: “Well, we should make four-year college free to everyone.” What if you focused more on two-year vocational education and you focused on making that free or low cost? Why are we putting everybody in the four-year college degree box? In some cases, that’s not only not good for them, but it’s actually not the right answer on numerous levels. The vocational education system—even defining what “vocational” is, at this point—needs to be reimagined. The power of these two-year programs and how people can slip in and slip out of them at numerous points in their career, that is a more important engine in a policy discussion for us to be having, whether or not four-year college is free for everyone.

Fuller: It’s playing out now. It’s visible now.

Anagnost: It is, and it’s not playing out in a very positive way.

Fuller: No.

Anagnost: All right? We need to have infrastructure that supports people in transitioning. It’s targeted. It’s targeted at industries, it’s targeted at skills that need to be reevaluated, and regions that need to be reevaluated. The world has a fundamental capacity problem, and it’s only going to get worse. Technology can help us evolve and do things more productively, do things with less negative impact on the world we live in. It’s proven that over and over again. It’s improved the quality of life for generations after generation. But the governments and the policy engines have to catch up to what’s happening.

Fuller: Very, very delicate challenge to balance the human impact of these things and also make sure that some of these exciting futures you’re talking about unfold as quickly as possible. Andrew, it’s been great having you here. Thanks for joining us on our podcast.

Anagnost: It’s been a pleasure. Thank you.

Fuller: Thank you for listening to this episode of the Managing the Future of Work podcast. To find out more about our project on the future of work, visit our website at hbs.edu/managing-the-future-of-work.