Listen to the episode or read the transcript below.

Transcript

Alan Murray: Leadership Next is powered by the folks at Deloitte who, like me, are exploring the changing rules of business leadership and how CEOs are navigating this change.

Welcome to Leadership Next, the podcast about the changing rules of business leadership. I’m Alan Murray.

Michal Lev-Ram: And I’m Michal Lev-Ram. Alan, this week we spoke to Andrew Anagnost, who’s the [president and] CEO of Autodesk, and he told us a lot about how Autodesk is using AI within lots of their products. No surprise AI was a big topic. And also how they see AI helping to build sustainable and affordable housing, which as much as we talk about AI, is kind of a unique lens, right?

Murray: Yeah, and I’ll tell you, Michal, what blew me away about this conversation, I mean, the AI stuff was very interesting, but what they were building some of that housing with was even more interesting. I never heard of it before. Industrial fungus. Fungus.

Lev-Ram: Fungus. Yes. I think both of us got a big kick out of the fungus piece, and we kept bringing it up.

Murray: We spend our life trying to make sure there’s no fungus in our house. Never thought. 

Lev-Ram: You’re not trying to build a house out of black mold is what you’re saying.

Murray: But this is a fascinating company. You know, been around for a very long time, got a lot of attention back when Carol Bartz was running it because she was so outspoken, had a unique way of talking about everything. It’s software that’s used to design, to build buildings, but all sorts of other things as well. And Andrew has a clear view of what its future should be.

Lev-Ram: Yeah, absolutely. He’s definitely been a big part of some of the more recent reinventions of the company. He did not drop F-bombs, so in that sense he is different from Carol Bartz. But he talked, like you said, about what he learned from his his predecessors and about AI regulation and a whole bunch of other things. So with that, let’s take a listen.

Murray: Andrew Anagnost, thanks so much for being with us. Let’s really start with the basics, because I don’t think we should assume all our listeners know. What does Autodesk do? Give us the elevator pitch.

Andrew Anagnost: Yeah. So, Alan, Autodesk makes software for people that make things, okay? And specifically designs and makes things. And, you know, just to kind of help you understand what that means, the act of designing something is deciding what its form fit and function is, how it looks, how it all fits together, how it actually works. Making something can be how you manufacture it, how you construct it, how you produce it if it’s a movie. So we make software for people that make anything. Building, bridges, cars, roads, rockets and special effects and movies, animations, all of that stuff. If they make it, they use our software to do it.

Lev-Ram: Okay, so sometimes we wait until towards the end to bring in AI, but we’re not going to wait. With this interview, we’re going to jump right in. So how has Autodesk been using AI? Because I know it’s not new to you. And what are the plans for the future?

Anagnost: Yeah, so AI has been something fundamentally we’ve been looking at for a long time. We actually coined the term generative design almost a decade ago. So it’s been something that we’ve been looking at in a disciplined way for quite some time. There is all types of AI. There is narrow AI, there’s general AI, and then of course they call it super AI. We’ve done a lot of narrow AI, which is focused a lot on insights, helping people glean information from a complex design that it would be hard for them to otherwise get. So a lot of the AI we’ve done has been some of this narrow AI specifically designed at helping people get information that they might not about. What’s the embedded carbon of the design? How many doors are there? How much should they cost? Using AI to kind of make those algorithms more fluid, help the user deal with a lot of other information. The more generative AI is things that help people generate geometry from constraints, like I want it to be constrained by certain design parameters, generate some geometry that satisfies those things. That’s some of the more advanced stuff we’re doing.

But if you look towards the future, what we’re going to try to do is take some pretty complicated tasks and really automate them with AI algorithms that allow people to spend a lot more time on the design problem, the design decision making, and ultimately, frankly, we’re going to change the way our customers work over a 10-year period. We’re going to have AI solutions that generate full kind of R1s of things 10 years out. And I think that’s an important paradigm change for some of our industries. They’re going to be looking much more at what am I trying to accomplish? Not how am I trying to model it in a computer.

Murray: It’s huge. But Andrew, you’re in the business of, if I can put it this way, precision AI.

Anagnost: Yes.

Murray: Your AI does math.

Anagnost: Yes, it does math.

Murray: We had an architect. I won’t use his name, but we had an architect at a design event we did in Macao at the end of last year. And he had this very cool tool where he could sketch just like three lines and the generative AI would turn it into a whole beautiful building design. And so I said to him, Well, is that architecturally sound? And he goes, Oh, no, no, no, no, not at all.

Anagnost: You couldn’t build off of any of that stuff. But it does help people explore ideas. But you’re absolutely right, Alan, we do math and we do try to understand not just what it looks like, but how it fits together, how you represent it, and how you can make really detailed decisions on it. You know, for instance, how energy efficient is a building? How leaky is it in terms of energy? You know, what’s its thermal efficiency? What’s the efficiency of its plumbing systems? Its HVAC systems? Its lighting systems? All the other systems that go in it? So you have to make really sophisticated judgment decisions and calls on this. So, yes, we do math, and these are complicated models.

Lev-Ram: On the creative side because you do math, but a lot of what your product is used for is very creative, right?

Anagnost: Yes.

Lev-Ram: From designing buildings to special effects, like you said.

Anagnost: Monsters.

Lev-Ram: Yeah. Yeah, totally. So on that on that front, you know, how far can this go and what do you see on the creative side? The design side? What’s the role of, you know, 10, 20 years out? What are humans still doing? How are we still in the picture?

Anagnost: Yeah. So we see a very collaborative future between AI and the human user. We do not see an adversarial relationship here. Does the AI ultimately put fewer people on each project? I guarantee you there will be fewer people on each project that builds or makes something in the world in the future. But we’ll be able to do more projects. And I think that matters over the next 20 years. Because let me kind of put it this way, okay? If we look at the world we live in right now, and I shared some of these stats with you before, but you look to the future. The world is going to be somewhere between eight and 10 billion people, depending on which estimate you take right now. If you look at that, we’re going to be adding anywhere between 35 and 45 Tokyos to the world every year between now and then. Okay. That’s a lot of new people. That’s a lot of new infrastructure. So fundamentally, the industries we serve have a massive capacity problem and that capacity problem’s connected to there’s not enough money, people or material to make and remake everything that needs to be made and remade without either destroying the planet or not being able to make it at all because you just don’t have the capacity. So we’re very much focused on that capacity problem.

Murray: Yeah. How do you do that? That’s a great place for us to dive into. We’re going to build 35 Tokyos. We don’t have the people, the materials, or the money to do it. How does a piece of software make all that more possible?

Anagnost: Yeah. So if you ever sat over the shoulders of someone using our software, what you’ll see is they’ve created a complete detailed representation of that entire thing they’re trying to build. Alright? Be it a skyscraper, be it a bridge, be it a car, be it, whatever it is, it’s a highly detailed representation. Inside that representation is all manner of good and bad decisions. The good decisions make it cheaper, use less material, and have a lower carbon footprint. The bad decisions make it more complicated, use the wrong materials that have massive carbon footprints, or drive cost and risk associated with the design up through the roof.

What we can do is we can manage a lot of information complexity behind the scenes and help them improve the quality of their decisions. And that’s the relationship that we see moving forward between the computer and our users is let’s help you make a better call based on information that you can get in real time that you could never process on your own. And that goes for the architect, the engineer, the construction professional. All of them are constantly making calls that increase either the cost complexity or carbon footprint of a design or a project. We want to make sure that those all turn in the right direction and we can provide that information for them. That’s how we can be kind of this collaborative co-creator with the various professionals in our industries.

Murray: Andrew, why don’t you tell us more about how that process can help us meet the huge climate challenge we have. We all know buildings contribute a great deal to massive amounts.

Anagnost: Yes. Okay.

Murray: How can you help control that?

Anagnost: So here’s a few of the things that we can help people do. All right. First off, you know that 30% of what you see in landfills comes from construction of all its kinds in buildings. All right. That’s just what shows up in landfills. So one of the things we can do is we can help people not waste things on a construction site. Construction, especially in building construction, is not a very precise process today. A lot of it really is kind of you got the drawings, you show up, even when you build a skyscraper, if you go there, you’re going to see a big pile of junk off to the side. It’s either going to be insulation materials, sometimes it’s steel, sometimes it’s wood, sometimes it’s other types of materials sitting there. We want to take that to zero. We want to create processes for construction that are more industrial, that mimic more the way a car factory. When you go outside a car factory, if you saw piles of stuff outside, the EPA is going to call and report that particular manufacturer for dumping. You don’t see that kind of waste. And one of the things that we can do for the construction industry is help them plan in a level of detail that prevents some of this overuse of material or waste of material. So that’s one thing that’s really out there that we can address.

The other thing we can do is we can help them know the impact of their choices. Every choice you make in a building has an associated embedded carbon number with it. Some materials are less carbon intensive than other materials. Concrete today is still very carbon intensive, but there are other alternatives and other choices that can be made. If we can help people make these choices in real time at the times of locking down design decisions, you change the carbon footprint of a building fairly dramatically.

The other piece is understanding the lifecycle impact of the building. How energy efficient is the building you’re designing, not how energy efficient do you think it is? How energy efficient is it actually? And that’s the kind of information we can actually give them. We can tell them precisely, hey, here’s how this building is going to behave in the real world. Here’s how your passive cooling and heating methods are going to work so we can make the lifetime look more efficient as well. But one of the big underpinnings here, and I really can’t say enough about it, is bringing the ideas of manufacturing in the industrial world to the world of making buildings and infrastructure. That kind of thought process. We’ll never get exactly there. But we need to embrace some of those ideas in the way we build buildings and bridges.

Lev-Ram: So you touched on this a bit already, but of course, on the other end of things, there’s this insane demand for building sort of tempered by sustainability or a need to be balanced with sustainability. But we had the CEO of Fannie Mae on recently and talked a lot about housing supply issues. And I’m going to just share this stat from Autodesk, from you guys that says that by 2030, 96,000 new affordable homes must be built every day. So I’m curious actually on a more global level what that looks like. And tell us a little bit more about just the role that you think Autodesk can play here.

Anagnost: Yeah. Michal, you’re getting exactly to the concept of why it’s so important to apply industrial processes to these things with all the housing that’s required. Because right now if we tried to make all those buildings with the current wasteful processes, we’re going to do some pretty damaging things with all of that housing that’s required. It’s going to be expensive as well. Now, by the way, just be perfectly clear, the cost of building a business isn’t all that goes into making a particular housing project expensive. There are other things that come in there. However, the cost of building something is a big driver. So when you have to drive down the cost and you have to minimize the amount of material used because there’s just not enough material to build all these things in a sustainable way. So if we can help people make repeatable choices that don’t make the building ugly, because this is the tension between the architecture profession and some of the things that we want to do with design technology is, the goal is not to standardize everything and make it look like the same thing over and over again. Heaven forbid. The goal is to help people minimize what they’re doing and make choices that actually minimize the amount of material or the type of material in really interesting ways and drive down the cost per unit. Because you’re right, with that kind of volume of building things, we just don’t have the capacity to do that. Look at the way we are right now. In the Bay Area where I live, we can’t build the housing and we can’t build it affordably fast enough. One of the reasons is that we build badly, but also we have zoning and permitting structures and all of these things that are complicated and get in the way of building the housing we need, not to mention other political issues. But one of them really is how you build. So we actually did a project with MBH Architects and an industrial manufacturing construction company called Factory OS for a West Oakland housing development that not only was cheaper and modually built, but it’s actually a beautiful building. And some of the sidings, some of the the cladding, the curtain walls are actually made out of mushrooms. Fungus.

Murray: Wow.

Anagnost: All right? Okay.

Lev-Ram: That’s very, very California.

Anagnost: What’s amazing about this is it’s a material that stores carbon, all right. Is very durable and can actually be shaped into anything you want it to be rather than using a plastic facade or something like that. It’s a beautiful material. It’s more durable than you think. It’s hard right now today to manufacture at scale, but it’s actually a wonderful material.

Murray: Andrew, can you eat it?

Anagnost: I wouldn’t recommend it. It’s an industrial fungus, but it is definitely a fungus.

Murray: We’ve been doing these for a long time. I do not believe I’ve ever heard those two words put together.

Lev-Ram: Industrial fungi.

Murray: Industrial fungus. I’m going to have to think about that one for a long time.

[Music starts.]

I’m here with Jason Girzadas, the CEO of Deloitte US, the sponsor of this podcast. Thanks for sponsoring it. Thanks for joining me, Jason.

Jason Girzadas: It’s a pleasure to be with you, Alan, and our privilege to sponsor this important podcast.

Murray: Well, it’s great to have you. This whole notion of generative AI has really exploded onto the scene and into our consciousness in the last year. It’s the fastest introduction of a new technology in history. How do business leaders deal with that and how do they separate the hype from the opportunity?

Girzadas: It’s a great question, Alan. The hype is real, but we also think the opportunity is more real and in fact an imperative for all businesses. The opportunity right now for businesses is around taking advantage of generative AI and other digital technologies for efficiency and productivity gains, with the belief they will continue to evolve and mature, such that there’s other opportunities for value creation and that new disruptions and innovations that we haven’t even seen the possibilities of. The challenge just to balance this opportunity. As a result, businesses have to diversify their approaches. It’s a CEO-level priority, an understanding of where and how these models are being put to use in your business operations. What are the controls put around data and data quality, as well as ensuring that the models are tested and actually validated like you would do any other customer-facing or highly sensitive system in an enterprise environment.

Murray: Jason, thanks for your perspective and thanks for sponsoring Leadership Next.

Girzadas: Thank you.

[Music ends.]

Murray: Let me change the topic on you a little bit, Andrew. You did an interesting interview with our Fortune colleague Diane Brady last year. She was working at a different publication at the time. But one of the things you said really caught my attention. You said that governments have a tendency to swing hard left or hard right, and you’re worried that they may try and slam the nail on AI and let other parts of the world run off with the innovation. How do you think we’re doing on that front?

Anagnost: I think we’re doing absolutely nothing. OK? And this is a problem here. Because, you know, I’m not anti-regulation. I think we need some regulation. We need rational regulation. Europe tends to over-regulate. The U.S. tends to under-regulate. And right now, we’re in a situation where if I could go a little philosophical on you for a moment. When Michal and I talked, I told her that I thought the advertising model was one of the worst things that ever happened to tech. And there was a real reason why I said that. And it’s important as we think about AI. When the advertising model started to create these tech giants and bigger isn’t always better, let’s be very clear here. Just because something creates a huge tech company doesn’t necessarily mean it’s awesome. Okay? When the ad model took over, it created a disconnect that for a long time hadn’t existed. It separated the needs of the end user from the needs of the customer. Most tech had been directly focused on the needs of the end users for its entire history. Then all of a sudden the customer was now the advertiser, not the end user.

So here we are. We’re heading into the world of AI. We already have this disconnect between the human using the technology and the person who is being the source of income for the technology. And now we’re going to layer AI on. And AI can either be a force for enhancing humanity or it can be disconnected from the goals of humanity. I’m very nervous in the structure we have around certain large tech companies right now that there’s going to be an increasing disalignment between the human-centered needs and the needs of the people that are funding the machine. And that is something, this is a classic place, a classic externality that government should be standing up to talk about.

Murray: Yeah.

Anagnost: And I have a particular point of view and I get my views from the MIT Initiative for the Digital Economy. There’s a scholar there called Sinan Aral where he kind of reminds people about what the government did with the telephone industry. I don’t know if you remember the time when you didn’t own your phone number. Maybe you do. There was a time when you did not own your phone number. The telephone company did.

Murray: Yep.

Anagnost: The FTC stepped in and said, Nope, this Baby Bell thing wasn’t working. Everybody owns their phone number. So now all of a sudden, all these Baby Bells are competing for us so that we don’t take our phone number away from them. A very simple piece of legislation that the government could put out there is you own your digital record, you own it, and we define the mechanism of portability. So now those companies who are monetizing your data in amazing ways, becoming trillion-dollar companies, they are beholden to you and your desires and your needs. I would really love to see something like that come from a government somewhere. Instead, we get very complex rules from Europe that I can’t understand and even my lawyer sometimes can’t explain them to me. And we get nothing from the U.S.

Murray: Yeah, yeah, we’ll put you in charge.

Anagnost: I think it’s dangerous. I think it’s a dangerous place to be when you’re dealing with a technology that really needs to be aligned with human requirements. And if it isn’t, you’re going to get what’s happened to my daughter’s generation where it’s going to take the anxiety of the cell phone generation to turn it into something far worse, in my opinion.

Lev-Ram: Are you optimistic at all? I mean, it doesn’t sound like it, but.

Anagnost: I am an eternal optimist. And I think, you know, the most effective way to be an optimist is to say what’s going to be great and to say what you need to make it great. And I think AI is going to be a positive enhancer to human productivity. I think it’s going to create massive collaboration and problem solving that we have never seen before. But we can’t let what happened with social media play out again. We can’t. It’s so obvious now. Did we learn nothing?

Lev-Ram: Okay, so we’ve talked about AI and the regulatory landscape and sustainability and industrial fungus, but we haven’t really talked about your own background. So tell us just briefly, what are the important milestones? What are the important aspects of your background that we need to know? And I’m curious where that eternal optimism comes from.

Anagnost: Yeah. All right. So milestones. Okay. You know, if we go back to the earliest milestones, I was let’s just say I was a problematic teenager. Briefly dropped out of high school, had lots of issues. Okay and really just kind of anti lots of stuff, right?

Lev-Ram: Was there fungus involved. Is that what you’re trying to tell us now?

Anagnost: Oh, you know, there could have been fungus involved, but yeah, I was much younger then, though. But we digress. My point was one of the things I learned from that period of my life is that, you know, the most important people in my life were not the ones that told me what I wanted to hear. They were the people that told me what I needed to hear. And I think, personally, for me, that has shaped a large aspect of my leadership philosophy. Throughout my entire career, I have always tried to focus on telling people a balance of what they want to hear, but also focusing really heavily on what they needed to hear. Because the people that told me what I wanted to hear were useless to me in my times of crisis. I’ve also kind of developed a philosophy that most people are capable of amazing things if they’re put into the right environment and given the right encouragement and place to do that. I was very fortunate in my early career to have been given a lot of mentorship and guidance that took me over the hump. I mean, I ended up from high school dropout to a Ph.D. in aeronautical engineering, computer science from Stanford. Okay. It’s not an easy journey. That’s not a straight line. That is a line with many squiggles. And and I’ve learned that along the way. There’s always somebody out there that wants to help you along that journey. So that was formative.

Lev-Ram: So I’m curious to hear where you stand on the skills versus college debate. And for you as a CEO today, are college degrees as important and necessary as they were before?

Anagnost: I think we need to look at education and skills development through a much broader lens than just a four-year degree. There are a whole set of jobs out there, especially in this new future of computer-assisted everything where people need to be comfortable using technology to solve a whole bunch of problems. And because the technology is going to be very knowledgeable about the specifics of the job, we need the people that are in those jobs to not only understand how to work with the technology, but also how to get the most out of the technology to solve a particular problem. Those people don’t necessarily need four-year degrees, and there’s going to be lots of jobs like that. And we are not well equipped right now to have excellent community college systems that create the skills that those people need, plus for degrees programs that create those other skills. And I think we do need to strike this balance in the future between people who have two-year degrees or credentials and people who have four-year degrees.

Murray: Yeah, you were a space guy for a while.

Anagnost: I was very much a space guy. I still am. Don’t ever assume you could take the space out of the space nerd. I left that industry disillusioned, actually went into software and tools. I got very passionate about tools and design capabilities and worked at a startup called EXA for five years and really got the software bug in a big way at that point. Really got focused on how I can create tools that help people design things.

When I came to Autodesk, I think this is an important piece of the puzzle here is, I came to Autodesk as kind of part of a rebel group. And this is one of the magical things about Autodesk, and we can talk about this a little bit, but Autodesk is we’re an 80s software company. How many 80s software companies can you name today that are still around? Okay, really, the odds are against you. There are not many 80s software companies that have managed to survive the way Autodesk has. And one of the reasons Autodesk is such a survivor is we constantly, constantly attack our own products and try to move our customers forward. So I was part of one of those attacks. I was part of trying to bring 3D modeling to the manufacturing industry inside of Autodesk. So I was I was a rebel, not part of the mainline Autodesk culture, an outsider. So I’m the classic kind of outside insider CEO.

Murray: You were hired to do that.

Anagnost: I was hired to be part of that effort. Yeah. Okay. And we were brought in. So I was an external force, for most of my career at Autodesk, I was never part of the main line. I wasn’t part of the hip crowd. I wasn’t one of the places where the big bucks were being generated. Ultimately, I came up to those ranks. But as Michal likes to say, I like to consider myself the product of two dysfunctional parents because I know they’re going to be listening. What I mean by that is that mom and dad didn’t always get along and they had very different points of view. One of them was Carol Bartz, the other one was Carl Bass, and I was, I think, incredibly fortunate to be exposed to those two people because they’re very different. They think very different things are important. But I got a lot of career advice from both of them.

Murray: Yeah. Carol Bartz, she was the first female CEO of a significant tech company, I believe. But those of us in the journalism business loved her because talk about somebody who spoke her mind, man. She was like the most candid person we had ever run into. A little bit profane, too.

Anagnost: Oh, yeah. I got I was on the end of lots of that profanity.

Murray: Frequent, frequent F-bombs.

Anagnost: Yeah. In my early career, I was on the wrong end of that profanity. And, you know, I have immense respect for the woman. You know, she really professionalized the company, verticalized it, had a very sales orientation. Carl, on the other hand, also very straight shooter. Very much a craftsman focused on the technology, the user experience, very different orientation. Those two orientations were just an amazing kind of sauce to be cooked in. And to see those things, combine that with my early background and the things I learned from both of them. I’ve tried to take the best of both of those mentors and turn it into something that creates a more lasting and stable company. I think I’m accomplishing that. But you know, we’ll never know until I’m done.

Lev-Ram: You didn’t have the right initials for the job, by the way.

Anagnost: I did not. A.A. does not work to get to the top of roll call in elementary school. But it doesn’t. It doesn’t get you.

Lev-Ram: Yeah, well, what are you taking a step back? What do you think your predecessors and you have in common? And what is it, like we already said Autodesk has been around for a long time and has survived kind of against all odds. What is it about the company and the leadership of the company that’s enabled it to reinvent itself multiple times?

Anagnost: So there’s a couple of things. First off, we are all incredibly passionate about the industries we serve. All of us are engineers. All of us are. There’s architects, engineers, special effects creators that work inside this company. We just love what our customers do. And a lot of us are steeped in what our customers do. And we have just a desire to make these industries better. And it’s really sincere. It’s at the core of the people that have led the company.

You know, when Carol came here, she came from the computer industry and, digital and places like that. I hope I didn’t quote that wrong, Carol. If I did, I apologize. And, you know, she left Autodesk saying, well, you know what, if God didn’t create it, one of my customers did it. And that ethos, that kind of idea of how passionate we are about what our customers do is core. Now, the other thing that’s core, and I think this is embodied even more so in kind of the things that I’ve been trying to do with the company is just, you know, let’s kill our own businesses, all right? Let’s be the first to kill our own businesses. Let’s not just think about where our customers are today. Let’s think about where they’re going to be tomorrow. So we took our customers from 2D drafting on AutoCAD to 3D modeling. We took them from 3D modeling to 3D modeling in the cloud. We took them from the cloud to design and make. We’re going to take them to the new age of AI, and we just keep looking where they need to go. And we’re not afraid. We’re not afraid to go after our own businesses. I always like to tell my salesforce, cannibals are much feared but rarely seen in the wild. And it’s true because everybody is afraid of the cannibals, but they never actually see them. And here’s the funny thing AutoCAD is less than 20% of our business now, but it’s bigger than it’s ever been. Yeah, we’ve been trying to kill it for decades because we’ve been trying to introduce new technologies. It replaces it, but we actually grew it. So where were the cannibals?

Murray: Andrew, thank you so much for taking the time to be with us on Leadership Next.

Lev-Ram: Thank you, Andrew.

Anagnost: Absolute pleasure. Thank you so much for having me.

Murray: Leadership Next is edited by Nicole Vergalla.

Michal Lev-Ram: Our executive producer is Chris Joslin.

Murray: Our theme is by Jason Snell.

Lev-Ram: Leadership Next is a production of Fortune Media.

Murray: Leadership Next episodes are produced by Fortune’s editorial team. The views and opinions expressed by podcast speakers and guests are solely their own and do not reflect the opinions of Deloitte or its personnel. Nor does Deloitte advocate or endorse any individuals or entities featured on the episodes.