Artificial intelligence (AI) and the Internet of Things (IoT) are here and they are changing the way businesses operate. Peggy Smedley and Ganesh Bell, president, Uptake, talk about what exactly they will look like for construction—and the drivers and big opportunities that exist. They also provide key advice for executives who are considering emerging technology, while also addressing the results AI can provide and the main risks at play.
If you want to listen to more recorded podcasts, click below to see the CONEXPO-CON/AGG archive of episodes.
Listen on your favorite app: Apple Podcasts | Google Podcasts | iHeartRadio | Spotify | Stitcher
Show Transcript:
Intro:
Welcome to CONEXPO-CON/AGG Radio, highlighting the latest construction technology and trends to drive your business forward.Coming up in March of 2020, CONEXPO-CON/AGG is North America's largest construction trade show. We bring you expert advice from your favorite brands, startups, and industry peers. For even more news, sign up for our weekly 365 e-newsletter at conexpoconagg.com/subscribe. We've got another great guest on the show today, so let's dig in.
Peggy Smedley:
Welcome to CONEXPO-CON/AGG Radio, brought to you by the Association of Equipment Manufacturers. I'm your host, Peggy Smedley. This month's episode is brought to you by CONEXPO Connect. Connect and win with CONEXPO Connect, the better way to research construction equipment. Enter for a chance to win a free trip to CONEXPO-CON/AGG 2020 by signing up for CONEXPO Connect by April 30th. Enter now at conexpoconnect.com.
Now let's dig in to today's topic, because it's a good one. This episode is all about artificial intelligence and the Internet of Things. AI is here, and it's already changing the way businesses operate. Consider this. Accenture says AI could double annual economic growth rates by 2035. Even more, it can increase labor productivity by as much as 40%. For construction, an industry that is facing a huge labor shortage, this type of technology is drastically needed.
So let's take a look about how artificial intelligence can provide some greater insight. AI offers intelligent data that enables machines to work and react like humans. It allows businesses to work more productively than they have in the past. We see industries like automotive, manufacturing, transportation, and logistics really embracing AI. But is construction falling behind? What do executives really need to know when it comes to artificial intelligence, and how will the technology redefine the way construction operates today?
To discuss all of this, we have invited Ganesh Bell, President of Uptake, to join the conversation. Ganesh, welcome to the show.
Ganesh Bell:
Thank you, Peggy. Thank you for having me on the show.
Peggy:
We are delighted. So let's get started. You know this space. You know the internet of things, AI really well. So for all of those contractors and the manufacturers who might be listening, let's help them understand really how is AI and IoT really helping shape construction, and how do you quantify what's at stake right now for everyone here who's really trying to help them understand how this is gonna improve their bottom lines? We've got a lot to talk about, but let's try to start right now from the very beginning if we could.
Ganesh:
Right. So at Uptake, what we do is help apply AI and IoT technologies to industries that are not being served very well with traditional software and unlock value from all of the operational data, and we call it industrial AI or IoT.
Construction is one of those industries that we're super excited about because if you think about this industry, it's one of the most digitally backward industries in terms of just in the last decade, we're kind of turning up the investments and embracing of digital technologies in construction. When I look at that, that's everything from around labor, around machines, around inventory, around project schedules, around managing projects. So there's a lot around machines and people, and our whole thesis is that in labor and in construction especially, we're sitting on mountains of data.
That can solve huge problems, productivity problems, efficiency problems that we can solve. That's really what we're focused on is working with customers to help understand everything from how the machines work and predictive maintenance around those machines and therefore once you optimize the machines, what can you optimize around the people and the parts and the inventory and the schedule, and therefore letting construction companies either improve their financial results, or either they save money or make money using IoT and AI.
Peggy:
It's important to understand the differences. I've been talking about this a lot, that there's a clear distinction you have to make when you're talking about AI, the internet of things, and even machine learning today. How do you help the construction industry understand where that distinction needs to be made, where it overlaps, and how they can actually get efficiency, profitability, safety, you talk about predictive analytics, all of that data that you just described that's so important to make them take that leap, because this is an industry, you know just as well as I do that this is traditionally an industry that's reticent to really embrace big emerging solutions. They always like to stay on the sidelines before they leap in. So how do we get this industry to really say, "This is where I need to go"? The big companies jump in, but the smaller companies are not so big to jump in right away. They want to watch and make sure that this is the right move for them.
Ganesh:
You're absolutely right. Any industry that's somewhat slow to embracing digital technologies, the one indicator that we look at is what's external interest in industries. The interesting data point there is we've seen the reports from McKinsey as well as from other places that say over the last decade, you get two different numbers, anywhere from 4 billion to $13 billion have been invested in software tech around construction. So that's a leading indicator, if you will, in terms of investments that are going into the sector because people see the opportunity in helping not just the big companies, but all the small and the medium companies.
To your point about what is the real value, I think if you demystify IoT, it's as simple as this, right? In all these industries in the past, we've used traditional enterprise software. The big companies use some ERP, work order management, asset management software and so on, but all of that software was designed for humans to enter data, whereas in the construction world, we have a lot of things that already generate data. I would argue, the most important data is all the machines generating data, all of the people at job sites and working in their operational data, it's the inspections, it's the work that people actually do.
So there's a lot of technologies now that take advantage of not just data that was entered in the old school ERP system, but all of the operational data coming from ... And all of this data has been around for a long time. The machines have been generating data for a long time. They've been sitting in history and databases and this is all sensor data. So when you think about all of the data from machines, people, and enterprises, weather data, pricing data, market data, transportation and logistic data, you are now coming up with a picture of all of the data in the world that informs the entire construction operation, and that's what we think of as IoT or internet of things. When all those pieces of data and sensors are being connected, then the entire value chain becomes [inaudible 00:08:01] and therefore now you can build applications on top that drive value.
When you think about all of that data now you have to provide insight, and a subset of AI, which is what most people in AI do, is machine learning. Our argument is that in all of these industries, especially in construction, the amount of data we're talking about is not human processable. So you really need machine learning technologies, whether it's doing predictive maintenance or labor optimization or parts and inventory optimization, any of these use cases, really needs to have machine learning and predictive analytics at the core.
So we try to simplify the technology with many customers and focus more on the business outcomes, because if you focus on the business outcomes, you'll learn to apply these technologies and not have a religious debate or even let's wait and see, because if you ask customers about, "Would you like to have less unplanned downtime? Would you like to optimize your labor force better? Would you like to have less inventory and optimize your parts?" Those are outcomes people can relate to and can attach financial metrics around, and then you drive those metrics. Therefore, the adoption of the technologies comes next.
Peggy:
A couple interesting points you just said there. One, I never thought that we'd talk about technology and religious debates in the same sentence. You made an interesting commentary there. But it was very interesting that you said about human processable, because there's something there that when you think about technology and you think about, again, IoT, AI, machine learning, you can't do it alone. Even as you guys look at what you're doing at Uptake, you have key partners that help you do it.
So when construction thinks about it, they have to look at this as saying, "We can't do it alone." So when you demystify that you just said earlier as well, the idea that I think construction needs to understand, construction companies need to learn about some of this to make it a little bit easier to digest or to embrace a little bit easier is that they're not going to have to do it alone. They work with a company like yourself to see that they can get it, but that you also work with partners, so all of this when they go to understand about all of this and they look that it's not human processable is that there's a lot of partners involved again that are helping digest all this data that you just described.
So I think that also makes it, because when we look at projects, there are a lot of things happening on the job site. There's a lot of people. So when we talk about gathering all this data, we're saying, "There's not all these people anymore. Technology's gonna take over jobs." Well, we're moving things around. So when we demystify that, isn't that also what you described? You have a lot of partners involved. There's a lot moving parts you have to explain to your customers, why the technology is so important.
Ganesh:
You're absolutely right. If you take, for example, an excavator, right, and it has a lot of data, I'm pretty sure in most companies and job sites you will have someone, let's say Sally who was the expert at knowing how to operate, how to maintain, how to know when to do the maintenance schedules on an excavator, because there could be a gear failure or engine failure at a job site and you don't want that.
In the past, there was always somebody like a Sally who understood that, but when you look at that same problem through the eyes of data and machine learning, you'll say that the excavator has so many sensors and so many generating gigabytes and terabytes of data.
Now, what if I can learn from all of the excavators? What if I can learn from all of the excavators of the same model across the entire globe? You can now produce better insight and make Sally more proactive, more efficient because she can then plan the maintenance window based on predictive maintenance because you actually have machine insight. That's not removing Sally from a job. It's actually making Sally way more productive, and that requires all that data to be processed.
For example, you're right. We have a huge partner ecosystem of people. AI augments people, that way we also augment other innovation in the world. All of our technologies run on the cloud, partners like Amazon, where we run our entire cloud infrastructure, or if it's new inspection in a construction site. We can partner with our partners like Kespry, who are an aerial drone intelligence software company, so they can go to the surveillance and put the image data and the aerial sensor data into our platform and we can build the analytics around it.
So it's actually taking all of the human expertise that exists in these industries. We've had people operate industries based on simple analytics and metrics, and if you talk to Sally about the excavator, I'm sure there's a lot of physics knowledge, manufacturing knowledge that goes into that equipment that informs her about when to do that predictive scheduled maintenance. Now she'll know more precisely from our machine learning analytics and our predictive maintenance software when to do that maintenance and when to plan the maintenance window around that.
So it's really enhancing the human productivity, but it's also enhancing safety of the human operator in those environments, so it does take an ecosystem. That's really why you need a fabric of software that we think of as industrial AI and IoT software that connects us. Once you solve the first problem, the next problem starts revealing so you can start following what we call the digital exhaust of the operation. So you can continually optimize further problems in the industry.
Peggy:
Do you find that you have to convince a lot of CEOs or board executives, or do they get it right away in the construction industry, or is it a whole process that you have to go through when we're talking about emerging solutions right now in the spaces that you're talking to?
Ganesh:
That's a great question. Even beyond the construction industry, we see this largely in the industrial world is that the few visionary leaders get it. They'll be honest. They'll say when they get it, they get the fact that they need to do something about it. They will be honest and tell you they don't understand what the next four or five steps looks like.
The most important thing I always talk about in this kind of embracing new technologies and what everybody would call a digital transformation or transforming industry through digital technologies, which is now happening faster because of ideas like machine learning, is that you need to have a believership at the top, the believership that technologies like machine learning, predictive maintenance, IoT will actually fundamentally reimagine your business. If you have the belief, it will lead you and guide you to make the right steps, and therefore surround yourself with the right people and partners.
So it's varied, but we see few people making those steps and many that have this as an agenda. To some extent, we also find people who are skeptical about it because they grew up in this industry and they would ask us about, "How can you help us? You're not a construction company."
We've seen this movie again and again in many industries. For example, we tell the story of when we first went into locomotives and rail, we didn't know a lot about locomotives and rail. But working with our customers, quantifying their subject matter expertise in machine learning, we're now able to help a company like Progress Rail do 90% of all the predictive maintenance using our software.
When we first went into the energy industry, we didn't know a whole lot, but then working quickly, this is the power of machine learning, is that we can quickly get to a point where we can help companies like Berkshire Hathaway Energy connect all of their equipment to us, wind turbine equipment, and help predict the maintenance.
In construction, for example, we already have over 200,000 assets, construction assets on our platform. In one case, we actually unlocked 23,000 technician hours with just one dealer in one year so these are real examples. Once people see the value, and the value being either they make money or save money or there's a business outcome ... We actually tend to talk about our, not software. We tend to talk more about outcomes. It's almost like outcome where is how we think about us, and that actually helps people build belief to get to the next step.
Peggy:
It's providing those richer insights. Since 2014, you guys really started doing all this, and you've had to really look at that. A lot of this experience you brought or created from the vision that you had being back from GE, helping Brad Keywell, who really started everything as the founder of Uptake, now you guys are looking at more industries, looking at ways to do it, but it's that experience you get from industry to industry, helps you gain new insights in what you can do as an industry like construction that says, "Look, I'm not comfortable with what I'm doing in construction," yet the larger companies see the need. They see that technology's not gonna slow down.
With all these buzzwords that get thrown around, you have to help an industry that says, "I want to put my toe in the water but yet I don't know if I want to put my whole foot in," so you have to help them see that predictive analytics is almost where they've got to go. They can't be reactive. They've got to be more predictive. They've got to be more prescriptive, and all these things that you do.
Now we're talking about deep learning and all these other things that go forward. Are they seeing that, "Look, we're gonna take you there, whether it's us or someone else, but we're the better choice"? Is that what you're really having to say to all these folks now as you talk to them more and more with the numbers you just said, with dealers and others saying, "It's us or someone, but you gotta go"?
Ganesh:
Yes. I think that's exactly it. The best conversations are with people who are gonna do this with or without us, right? It's almost like we advise people that, "You should do this regardless of whether Uptake is doing it or 10 other companies are doing it."
If that's a belief that you can build for yourself independently, then it's a good conversation, because then we can help you accelerate that. Then we can help you clarify why you need to do something or partner with a company like Uptake because we can derisk your roadmap or we can help you with a better journey map. It's absolutely that level of conversation.
Growing up in Silicon Valley and looking at a lot of these kind of problems, you know a market is real when people try to scratch their own itch. They try to self-service themselves. When I look at all the construction companies in the last decade embracing digital technologies on many fronts, it actually tells me that this not an old school idea. They're trying to transform their own industry.
An example for me is Kiewit, a big construction company started buying software companies, buying software companies, because they were tired of old school vanilla enterprise software companies like SAP not serving those industries. So they wanted to build their own construction software.
That's a healthy bet. Now, you can argue whether that's gonna be successful or not, but that pattern is important, right? You need dedicated industry purpose build software, but to your point, that also can be informed by expertise in other industries, which is how we can help our customers.
For example, in construction, you have fleets of trucks. Now, because we've done fleets of railroad, locomotives, and we can now translate that to fleets of trucks that we're helping with many customers around the world or fleet of fighting vehicles for the Army, we can bring a lot of those cross-industry learnings and derisk a project within construction, for example. That cross-industry learning and building the AI and machine learning engines and the algorithms and the models on top and training them on a billion and a half hours of learning data means we can deliver value quickly.
So we tell most of our prospective customers that, "You have to do this, but you've gotta do this in partnership with companies like Amazon, companies like us to be able to build your dedicated solution to transform your industry."
Peggy:
So companies like you're describing like Kiewit, when they find success and that there's gonna be those that are not finding success. What happens when partnerships fail and they go forward? 'Cause some are very successful. We're seeing very good success partnerships and other ones don't. What happens when that goes forward?
Ganesh:
So the key thing in digital is about if you learn anything from the world of software, it's about iterations and minimum viable product. That means that one or two steps may fail, but your breakthrough is you're coming up with ideas and archetypes that will not fail down to execution. We could have a debate about is Uber going to be successful or Lyft going to be successful, but one thing I'm sure we'll not have a debate about is ride sharing here to stay, right?
So the best way for companies to embrace this is to derisk with partners, build a belief system around digital, know the investments and iterations that they have to make, and fail fast. This is the idea of the lean startup where you build minimum viable products and quickly do it, and then there's gonna be a huge talent gap in the industry, right?
At the end of the day, it comes down to a human-driven transformation of why would some of the best software technologies come and work construction? Well, they could. If the purpose is driven in the right way and you provide the right platform and infrastructure, you can actually attract great talent.
That's what we're seeing now is in many industries, we're actually able to attract great talent because they want to solve problems across all of these different industries. The smart leaders in many industries including construction are making their bets with the right partners and also looking at their own digital transformation agenda internally and looking at the talent they need for the next two and three years of this transformation.
Peggy:
When we talk about emerging technologies and we talk about trying to make investments in IoT, AI, machine learning, are we not having the right conversations by telling companies that they are going to experience some failures? Isn't that a good discussion to have? Because if we tell them they're always gonna succeed, isn't that the wrong discussion to have?
Ganesh:
Absolutely. I think you're absolutely right. In some industries, that can be scary, right, especially-
Peggy:
Because we're talking people and processes, right? Technology isn't just technology. It's about the people and the processes, right?
Ganesh:
Exactly. Failing fast may be an acceptable terminology in a software company but may not be acceptable in an industrial company, right, a construction company. Especially with the financial crisis 10 years ago, right, construction companies are looking to protect their margins and relieve their costs.
Peggy:
Well, they don't have many margins right now. They're very thin. They've always been thin, right?
Ganesh:
Exactly, so they have to drive productivity. They're caught between doing the cost saving initiatives and investing in new technologies for growth. Actually, they're not mutually exclusive. We almost find that if people focus on financial outcomes, you can actually unlock dollars from your operations and invest in innovation, but they have to go hand-in-hand. That tends to be somewhat bigger.
The main thing both Brad and I have seen across multiple industries having done this for a good part of our last decade is that we see that the winners invest big and go big. Big doesn't mean a big failure. Big means a commitment, but within the big bag there are many small initiatives, many small quick wins, many small areas where they can quickly experiment true value, build traction.
Sometimes the idea could be really simple. For example, predictive maintenance, we know we've been doing it for quite some time and it's a no brainer in many cases, but you still have to build a human confidence factor in the predictive maintenance, right? So for that, we learned that in the early days that we can give the best predictions, but if it's not accepted and used, then it's useless to drive the value.
So we've built things like what we call a model performance report of our AI predictions, so people actually understand the confidence factor and how the predictions are being accepted, so it's about providing radical transparency. So humans build confidence in the predictions that the algorithms are giving them and when they accept it, they're able to trace it back to a financial outcome or an operational outcome which builds confidence, which builds their ability to define the next investment target or next use case within their own enterprise. So it is cyclical, and you're absolutely right.
Peggy:
So when you think about that just in general, maybe the question should be, they have to be thinking about technology now. There's no other alternative, though, but if they're not ready, that could be also a negative. So what advice do you give companies right now? What's your best advice for them today in this environment?
Ganesh:
So I think you brought two archetypes of companies, right? Companies that believe that technology is going to be a big productivity lever or help them unlock dollars from their operations to invest in innovation. If they have the belief, it's okay if they don't have a clear roadmap. It's good enough to start with a belief and partner with companies to go build a journey map, if you will, on how they will drive digital transformation.
If a company doesn't have that belief, it's still okay. All they have to do is understand the business outcomes they're looking for and allow for companies like us to come and show them how they can accomplish a business outcome. They shouldn't have to worry about cloud versus AI versus machine learning versus IoT-
Peggy:
Or worry about the edge. [crosstalk 00:27:30] those little things, right, 'cause that's what they hear. Those are all those buzzwords, right? They get lost in that.
Ganesh:
Yeah, yeah. So if they focus on the business outcomes, then that's one of the things we've learned in our journey over the last four and a half years, is that understanding our customers' business outcomes and driving for the business outcomes versus technology, so that's why we're engaging with CEOs, CFOs, and bridging to the CIO, because the CIO, she may have a great idea on how to transform their business. Sometimes they don't have a seat at the table and we can help them.
So it's really focusing on the outcomes, and that's why I use the word ... Sometimes if you start purely with technology and not know the exact outcomes you can impact, you get into these debates of cloud versus no cloud versus traditional analytics versus machine learning analytics, physics-based analytics versus data science and so on and so forth, and you end up in, "Prove it to me," and trials and POCs and most people who have done data science will tell you, people end up in the trial hell of doing constant proof of concepts and never seeing anything to production because you are not driven by a business outcome.
Peggy:
Well, I have to tell you, Ganesh Bell, President of Uptake, it's been great to have you on the show today. Thank you so much.
Ganesh:
Thank you, Peggy.
Peggy:
All right. As you've all heard here today, there is a big opportunity to leverage AI to help power the construction industry. Construction can truly tap into the data and the analytics to understand patterns, even automate tasks. If you look at the bigger picture, I think you have an opportunity to cut costs, heighten productivity, and even enable preventative maintenance and so much more if you look at all the great opportunities that lie ahead. Clearly it's coming. Technology and AI are blending together and it's time construction companies prepare for that next evolution. I'm your host, Peggy Smedley.
Outro:
And that's going to wrap up this edition of CONEXPO-CON/AGG Radio. If you liked the show and think other people should listen too, make sure to subscribe and maybe leave a review on iTunes. We'll be back next time with another great guest. Until that time, be sure to visit conexpoconagg.com/subscribe to sign up for our weekly e-newsletter. More than 30,000 other construction industry pros are already receiving news and insights to move their business forward.
