Bricks, Mortar, And AI: Could ChatGPT Be The Key To Unlocking New Efficiencies In Construction? With Robert Salvador | Ep. 244

COGE 244 | ChatGPT In Construction


Artificial Intelligence has become a powerful ally to many industries, and the construction business is no exception. It has become a weapon that conquers inefficiencies and unlocks new business possibilities. In this episode, Robert Salvador, the CEO of DigiBuild Software, explores the value of AI and navigates the ins and outs of ChatGPT in construction. Emphasizing the significance of AI in the construction business, he explains how you can practically leverage it. Robert also provides concrete steps to make this technology work for your company. Let’s traverse into the exciting landscape of AI Construction with Robert’s expertise today!

Watch the episode here


Listen to the podcast here


Bricks, Mortar, And AI: Could ChatGPT Be The Key To Unlocking New Efficiencies In Construction? With Robert Salvador

We have a unique discussion lined up for you. Our guest is Robert Salvador, the CEO of DigiBuild, a front-runner in digital construction technology. We will be exploring the realm of artificial intelligence, focusing on a tool known as ChatGPT. It’s not just for your kids writing their papers. It is a technology that you can use in your construction business. Our conversation is going to cover the ins and outs of ChatGPT, what it is, its significance and how to leverage it in practical ways from business development to estimating, pre-construction, project management and a variety of different use cases that we’re going to discuss.

We’re not going to shy away from its limitations. We’re aiming to provide you with an introductory understanding of ChatGPT and going to give you some concrete steps at the end to make this technology work for your company. With Robert’s expertise, we’re set to traverse the exciting landscape of AI in construction. Let’s dive into this fascinating conversation.

Robert, welcome back to the show.

Eric, I’m happy to be back. Thanks for having me.

It’s my pleasure. We’re going to dive right in and talk about ChatGPT. Before the audience goes away, don’t worry. This is going to be very relevant to the construction industry. I want to dive right in. Treat me like a fifth grader who knows nothing about ChatGPT. Tell me, please. What is ChatGPT?

ChatGPT is what’s called an LLM or a Large Language Model. It is probably the most sophisticated one that’s ever come out. There’s a company called OpenAI. It’s probably the biggest and most technically-advanced AI company out there. ChatGPT is a front-end user interface or a chatbox where you can interact with the underlying large language model. That means you can chat with artificial intelligence. Artificial intelligence has the ability to recognize your speech patterns.

Normally, to engage with any artificial intelligence, it takes a bunch of very complicated knowledge and the ability to code and do all these different things. What makes ChatGPT special is the fact that it can understand you and talk to you. It allows unsophisticated users to harness the power of artificial intelligence. ChatGPT is this artificial intelligence that came out that individuals are using for anything and everything, things in construction and their college papers. This artificial intelligence is writing papers for people. We will get into the use cases of what ChatGPT and AI in general can do in construction. It is a breakthrough in artificial intelligence that we haven’t seen in our lifetimes.

It’s important because you mentioned people writing papers. My kids use ChatGPT. At least they say they don’t but I know they do. We have heard about all those use cases but I’m a construction executive. I run a construction company. Why is ChatGPT important for construction?

In our industry, there are so many different workflows and things that I would consider time sucks or things that take away our project team’s bandwidth. A typical construction project or a large one will create over 700,000 documents. Doing all the things you have to do to manage those, emailing suppliers, looking on the internet for substitutions for certain material, trying to locate logistics or lead time information and crunching all that information is very difficult.

As people, we have a limited ability to gather that data. You can only google so many things in 30 minutes. What artificial intelligence can do is gather all that data, look at all that data and pattern match all of that data in a split second. We will talk about what we do at DigiBuild. Think about a situation where you’re trying to get a quote from a material supplier. Instead of you having to manually send an email and make a phone call, which might take you 15 minutes and you get contact with 1 or 2 suppliers, at the click of a button, artificial intelligence can reach out to 100 suppliers and do it in an intelligent way. These are learning models.

The artificial intelligence can say, “I’ve read all about these building material supplies and all these supply chain information and specifications. As the artificial intelligence, I understand it.” Before artificial intelligence and things like ChatGPT, there were tools out there but you can’t go into your CRM and say, “Reach out with a custom message to all these different people in a different way.” The automation is not there. The software doesn’t understand you. It’s a piece of dead software whereas with artificial intelligence, the software learns.

You say, “This is Type X drywall. These are the vendors that typically provide Type X drywall. This is what Type X drywall does on a construction project.” Artificial intelligence starts to learn that. It can scan the internet and say, “Type X drywall. I know Georgia-Pacific makes that. Did you know that Armstrong also makes that too?” It can start telling you intelligent insights like, “Builder, instead of getting your Type X drywall from Georgia-Pacific, you can get an equivalent from Armstrong for 10% less.”

The real magic in what artificial intelligence can do in construction is it can learn all of these industry-specific pieces of knowledge and information and apply that to our industry whereas before, you had to build a specific tool for a specific thing. If you feed artificial intelligence enough of this data, it can learn by itself. It can start to say, “I suggest this substitution.”

It can look at 100,000 different data points that would take human years or a long time to look at that same data. If you apply that to all the different things that we do in construction and construction management, you can see all of these pieces of busy work, the office work, the project manager work and the project engineer work. You can make that much more efficient and effective. That’s why people are excited about ChatGPT in construction.

Let’s explore this a little bit more. In interacting with ChatGPT, there are a couple of issues that people have. The first one is, “How do I know that the answers are correct?” The second one is, “How do I develop prompts that get the information that I want?” In my experience, the quality of the prompt drives much of the quality of the information. First, speak to the idea of, “How do I even know that this information is correct,” and the quality of the prompts, please.

I’ll take that in reverse because one plays into the other. Let’s talk about the quality of the prompt first because that will play into the effectiveness of the information that you get or how accurate it is. The quality of the prompt is everything. If you’re not that familiar with artificial intelligence, it’s easy to think this is some human on the other side who thinks and has cognition in the same way you and I do. That isn’t the case. Artificial intelligence is pattern matching.

Without going too deep, at the end of the day, if you want to think about it and abstract it far back, there is insane pattern matching that happens with artificial intelligence. It’s taking data from this side of the internet and putting it next to this side of the internet. All of a sudden, you have a bunch of information that you wouldn’t have found before. Prompting it properly is everything because the AI doesn’t understand. It doesn’t reason as you and I do. It pattern matches. First and foremost, there are tons of good resources out there on how to prompt properly. If you’re going to use ChatGPT, you should look at that because you want to be able to do it the right way.

One of the ways that is a good practice is to start small and then add more. One example of the way we use it in construction and someone can do this at home, is you can go into ChatGPT and tell ChatGPT how you want it to act and who you want it to pretend to be. You go in there and say, “ChatGPT, I want you to act like a construction manager and procurement manager with twenty years of experience on big commercial construction projects, managing supply chain, building materials and all these different things.”

You are putting that in the prompt at the very beginning.

That locks in ChatGPT to certain pieces of information on the internet. ChatGPT is only looking for words that have building material, supply chain and all of those things because you don’t want ChatGPT looking at tennis. You don’t need that information. You want building material and supply chain information. Prompting it upfront to focus on that puts it in that ballpark or range of information you’re going to be looking for.

How narrow should you be in the parameters that you give ChatGPT?

You want to start broad and then drill down as much as it makes sense. Start with the industry. We tell ChatGPT, “I want you to act as an expert in construction management and supply chain.” Let’s get some substitutions for building materials. You might say, “More specifically, ChatGPT, I want you to focus on drywall and Division Nine workflows in construction.” You can tee up some specific areas where ChatGPT might look for good information. I would say, “ChatGPT, check Engineering News-Record as one source of information, BuiltWorlds and Construction Genius Podcast.” These are some examples of places you might find good information.

I would say, “ChatGPT, tell me you understand this and then hold for more instructions.” ChatGPT would say, “I got it. I’m starting to look at these certain areas. I’m waiting for more instructions.” I would drill down more and say, “Here’s what I’m looking for and how I want to get it. ChatGPT, I’m looking for a type of drywall that is a bit rarer. It’s called Type C drywall. I want to know in my commercial construction project some other types of material that I can use that have similar qualities as Type C drywall. Go out and find me some vendors and/or material types that might meet the specifications that I need.”

We’re three rows deep. We told them who you are and where they should look for something. We told ChatGPT exactly what I’m looking for more specifically. ChatGPT would return ten different suppliers or suggestions for alternative substitutions or material types for that material. When you get that information back, you can drill down even more, “ChatGPT, you told me Type X drywall. I asked for Type C. Let’s try that again.” It’s also important to use your expertise to make sure that ChatGPT is providing relevant information but once it starts to provide that info, you can specify more.

That’s the thing about an LLM or a learning model because it’s going to learn as you tell it more things. AI goes, “I messed that up. You told me why I messed it up. I understand more and I won’t mess it up next time.” At DigiBuild, we do it in a more automated way but you can do this through the chatbox with ChatGPT. By the end of this, it’s given you ten different alternatives or substitutions. We’re going into the second part of your question, which is, “How do you know this information is correct?”

You can take these 10 vendors and spend the extra 5 minutes to google them to make sure they’re relevant vendors of Type C drywall or call it yourself. That’s where a lot of people are worried about artificial intelligence replacing jobs. Artificial intelligence will empower people to do their jobs better. Those project managers aren’t getting replaced. Instead, ChatGPT did all the hard work. They can do a quick Google search to make sure all that information is relevant. We like to say at DigiBuild that we use AI to empower, not to replace.

It’s important for people to understand that because I can see someone who runs a construction company saying, “Let me give ChatGPT to the marketing intern to take a look at. Maybe we will come up with something.” The marketing intern may know how to use it but they don’t have the years of experience to make it useful. If you want to exploit ChatGPT, it’s going to be based on the quality of the prompts that you give it and the sophistication of the person using the tool in terms of being able to filter the information that it brings back, pick out what’s right and discard what isn’t right.

Realize that ChatGPT is very smart, even if the person using it isn’t good with technology as long as their information is good. Think of a superintendent in the field. They’re using the ChatGPT app, which you would never expect but they might not be great with prompts yet or tech but they know the information and their material like no other. ChatGPT is good at deciphering some of that. Even if the end-user is imperfect as long as the information is somewhat close to what it should be, you can help it to find what you’re looking for.

Let’s take a use-case scenario. Let’s say I have a specific project. I’ve handed it off to the project manager. This project manager is sophisticated. They’re able to interact with the technology and all that stuff. Can I set up ChatGPT to help me run the project through the whole cycle of the project? If that’s the case, how would I do that?

The way we look at things like artificial intelligence in the future of construction is AI will be a digital project manager that helps your real project managers. Think of it as having an angel on your shoulder when you’re managing a job where you can say, “I don’t have time to do this. Digital project manager, do this for me or get me this information quickly so I don’t have to waste my time finding it.” To your question, part of what we are doing at DigiBuild with the supply chain side of things is eventually, these models will be trained specifically on construction. If we give all of these millions of pieces of data to ChatGPT, you have to realize the sophistication of this.

You can upload specs, blueprints and things that if you copy and pasted anywhere else in the world, the software wouldn’t understand. Artificial intelligence understands it. I’ve taken email strings in construction with no prompt at all other than to say, “Help me decipher this email string.” It’s an email string between a GC and a sub. Those can be big arguments and all that. Copy and paste it into ChatGPT. It was able to differentiate and tell me what happened in that entire conversation, “It looks like General Contractor A is claiming that Subcontractor B is change order is not valid because of this contractual contract clause. Our recommendation is for Subcontractor B to respond in this way.”

This is fundamental in terms of how you can use this in practical ways. You’re in a pinch and bind. You’ve got that email chain that you’ve described. I’m a project manager. I’m going into a negotiation for a change order. He or she said all the details. Take that email chain and throw it into ChatGPT, “Give me the top ten takeaways from this email chain.” Give it the prompt, “I need to prepare for a change order negotiation. Please, give me my three strongest points and the three strongest rebuttals that the person I’m negotiating with is going to come back with. I can roleplay with my team and get ready for that change order negotiation.”

You can put your contract into the language model too. You can combine the email string I sent you with, “Here’s our contract with the GC.” You don’t want AI to do legal work for you. You still need a lawyer and all this but as a filter, it can say, “Based on that email string, clause number 3.21 was breached. They didn’t give you a proper 48-hour notice before terminating you.”

If you think about the fundamentals of how that is valuable, imagine you didn’t have ChatGPT. How would you do this in a general contractor’s organization or a subcontractor? 5 of you from the project plus your legal team and a VP would have to sit around a table for 3 hours and eventually, after 3 hours, you would get half of the information that ChatGPT gave you in a second.

Going back to your point, you can train ChatGPT on construction-specific workflows where throughout the course of the project, it can be suggesting things like this and say, “Based on this schedule you gave me, it looks like your plumber lags behind every three months. Here’s a suggestion for changing your schedule to make it more optimal.” Put in a subcontractor’s prequalification information and then go, “ChatGPT, based on these lists and places on the internet that show good subcontractors, give me a compare and contrast of this subcontractor that we’re prequalifying with a top-tier subcontractor out there like Anning-Johnson or something.”

You can train ChatGPT on construction-specific workflows. Click To Tweet

To your point, that is how you can train these models to recognize and understand construction workflows. At DigiBuild, we specifically do that in the supply chain. We have the largest material and material supplier database in the industry. Think about hundreds of thousands of suppliers, material types and all this. AI can crunch that data around pricing, lead time, substitution and schedule clashes much more effectively and efficiently than any human can do on their own. When you turn that into a product for the construction industry, it ends up being so powerful. That’s what we do for our customers.

Let me ask you. One of the things that is exponentially increasing in the ChatGPT world is the use of plugins. Can you explain that a little bit to the audience? There’s the basic ChatGPT interface but then there’s this plugin aspect to it as well. Can you speak to that a little bit, please?

ChatGPT is a language model. It’s not understanding logic and cognition the way you and I do. It’s matching words together. As smart as it seems and it is incredibly smart, it is pattern-matching words. It’s limited by the words. I can’t upload a video or some media into ChatGPT but with the plugins that are coming out, I can upload an image and ChatGPT can interact with it.

COGE 244 | ChatGPT In Construction
ChatGPT In Construction: ChatGPT is a language model. It does not understand logic and cognition the way people do.


There’s one that’s come out called AutoGPT. If you’ve seen the movie The Matrix, you remember there are those agents or the bad guys. They’re a computer program. Their only goal is to destroy Neo, the main character of The Matrix. They have these AutoGPT agents. You create these agents and give them a goal. They don’t stop until they reach that goal.

One of the plugins available is you could create this agent and say, “I want every construction contractor’s social media information so I can get to them on social media.” This artificial intelligence agent will start doing things and will not stop until it gets you all of that information. It will automatically scrape the internet for all of that data and do all these different things.

There is a ton of plugins that allow you to do much more with artificial intelligence. It allows you to make AI movies. You may have heard of the popular rapper Drake. There’s an AI Drake song that came out. It’s called Winters Cold. If you listen to this, you’re like, “That’s Drake. That is the rapper.” It’s not. It was made by artificial intelligence. In the same way in construction, we integrate with other software. Procore integrates with DigiBuild. You’re starting to get these integrations where you can build on top of this powerful AI model.

That’s what will allow you to do things like build these agents specifically for construction. Imagine building an agent saying, “Do not stop until you get me the world’s best price on blowing insulation.” This thing will reach out to people, do all these different things and won’t stop until it has that information. The AI is getting started. Some of these plugins are starting to show incredible potential in many different use cases of AI in these language models.

Are you at DigiBuild thinking of building a plugin for what you offer? Is that not necessary?

We work through a plugin on top of ChatGPT and integrate it with ChatGPT. There are two ways to use ChatGPT. One is talking to it. We’re talking about prompting it on your own. The other way is training a model on top of ChatGPT, which is much more difficult but it’s much more powerful. We are training a model. Every day, we are plugging in thousands of different material types and different situations on a construction site. Slowly, this model is learning.

I’ll give you a specific example. We located these Google Nest Thermostats. This project was specked with these Google Nest Thermostats. They don’t make them anymore. We plugged this into ChatGPT. Before, a couple of us would have to go online, look for substitutions and see what’s equivalent. Now that ChatGPT is learning all this, instantly it goes, “Here are five equivalent material types that you can take to the architect and tell them you can order this instead. It’s cheaper and it will get there faster too.”

The point is we are building on top of some of these APIs and API integrations. This is foundational. This isn’t some cool tool. Bill Gates said when ChatGPT first came out that this isn’t just some amazing tool. This is a foundational change in the way business will get done because there’s no way to beat this. Spreadsheets versus AI, there’s no way to win. We’re not going to get rid of our employees in construction but this will empower everyone to do things in such a more efficient way that eventually, it will be a big disadvantage if you’re not using AI.

We're not going to get rid of our employees in construction, but AI will empower everyone to do things in a more efficient way. It will eventually be a big disadvantage if you're not using AI. Click To Tweet

Let’s say I’m a large general contractor and I want to codify in an accessible way the way we run projects and the type of subs that we want to work with it. Do you see someone externally but also someone in-house perhaps doing that model training on a particular process or project so that you are tapping into the AI and using it to make things more efficient and productive?

The use cases of AI span the construction project lifecycle from pre-con to early development to the course of construction and even to turnover in your specific scenario there. What’s prequalification for most companies? There are 1 or 2 people in a risk department. They send you out this piece of paper that says, “Give me all of your basic information, references and past job history.” If they’re sophisticated, they ask for your finances. 1 or 2 people in the risk department go through the paperwork and look for all these things.

Eventually, you could plug 100 of those prequals into ChatGPT and train the model. When you get a prequal in, ChatGPT looks at it and goes, “Section 7, they don’t look strong enough on their finances. Section 9, the biggest project they have done is only 100,000 square feet. The project they’re bidding it’s 250,000 square feet. That makes them risky.” There comes a point where it’s like a fantasy football draft. You’re the user. You’re still drafting your team but you have all this information in front of you on a very easy screen and it’s already been classified and told you what’s relevant.

When you draft a fantasy football player, it goes, “Patrick Mahomes is great at throwing touchdowns. If you’re looking for someone who throws a lot of touchdowns, that’s who you should draft.” It’s the same thing with a subcontractor, “ChatGPT has identified that this one is high risk but they’re low price.” You as the project manager can decide if you want to draft them or not.

Think about that at scale in a construction company, both the savings and cost, staff time and then risks because you standardized it and then you get both the benefits of artificial intelligence looking at it and the benefits of a human construction manager or a risk manager looking at it. Things like that across the office or the spectrum will be made better by AI and ChatGPT.

Let’s dive a little bit more into a use case to roleplay it. Let’s say I’m a hard-bid heavy civil contractor. I have to make sure that my takeoff is on point when I’m doing some estimating and all that good stuff. How would I use ChatGPT either now or in the near future in terms of a scenario like that?

First and foremost, let’s assume that you’ve already trained a model a little bit and you’ve been using ChatGPT a little bit. You’re not brand new. As you are estimating even pre-con, you can use ChatGPT to look at the risks and compare them to previous projects. ChatGPT can say, “This is a very similar project to what you did last year on this job. You went 10% over budget on this line item. Make sure you account for that on this future project.” We can use past data effectively to say, “Here’s a risk.”

You can get into pre-construction, “ChatGPT, here’s our material quantity takeoff. Here’s what we need. Go to 100 suppliers and get me pricing from all of them. When you get that pricing back, I want you to do your best to drive down all of their pricing at least 10% more.” With the click of a button, ChatGPT can do that. It can reach out to 100 material suppliers and even respond. The material supplier goes, “Hello, DigiBuild.” They think they’re talking to a person. It’s ChatGPT they’re talking to, “Can you specify something?” ChatGPT might be able to do it. Maybe you need a project manager to help but either way.

With the click of a button, you’ve got 100 quotes. You’ve driven down that pricing. You have the guaranteed best price. What ChatGPT can do is identify bulk buying opportunities. It might say, “You might not realize it but you’re doing three pavement projects within 50 miles of each other. Did you know that if you bulk-bought all the material from this manufacturer or wholesaler, you would save 40%?” Right there, you’ve got all your best material pricing.

ChatGPT can say, “Give me your job schedule. I have your job schedule. I have the lead times on these materials because we got quotes from them. We can tell you when you need to order all of this material.” Let’s say it’s a 120-day lead time, “Make sure you order that by May 1st to get it there in August.” It can notify you, “It’s May 1st. Order your material.” I’m talking a lot about it but this is all 1 click or 2 clicks of a button. You’ve got all of your risk from previous projects. You’ve gotten the best material pricing.

You know when you need to order this stuff by, “One of these materials is no longer available. It’s a 200-week lead time. I can go in there and ask for a substitution. I’m screwed. What are some options that I can do?” ChatGPT will say, “Try this equivalent from whatever brand.” You go through the course of construction order your material, get all this data and crunch it. You have insights for your next project. It gets much bigger than that but that’s a quick and easy way that a civil contractor could be using it.

Let’s talk real quick about some of the limitations of ChatGPT because it’s important for people to understand that this is not like NFTs. This is stuff that you can use in your business. It’s not hype and it’s not going away. Tell us the limitations at least at the moment or perhaps even long term when it comes to using ChatGPT or similar types of tools.

First and foremost, I’m sure you’ve seen the meme that’s making its way around the internet, LinkedIn and Twitter. It’s a big building and on the side of it, it says, “ChatGPT cannot finish this building or build this building.” That’s true. The first limitation is this is data. It’s data extrapolation. It cannot build your building for you. Anything that’s in the physical world, it cannot do for you. It can help you with the digital elements that go along with that physical world but that’s the obvious part. Maybe in the future, we put artificial intelligence in robots and then maybe things change but there is a clear distinction between the physical world and the data world.

COGE 244 | ChatGPT In Construction
ChatGPT In Construction: AI cannot build your building for you.


The second thing is ChatGPT is still insulated to its program, meaning when I say ChatGPT can send an email for you, yes, if you build out a plugin to an email service where it can send those emails. ChatGPT by itself cannot access Gmail or upload computer code to GitHub. It’s still a program that’s insulated by itself and you have to connect it to other things to get it to do more than give you information. There is still some technical sophistication that has to be there to connect it to these other programs to make it more useful.

Another one is making sure that information is correct. There are a lot of things on the internet that are not correct. ChatGPT is pattern matching. It doesn’t know if it picks up some information that is not valid or accurate. Maybe it pulled something from a bad website that has bad information. The info isn’t always accurate. You may have heard that there are some things called hallucinations in artificial intelligence where the AI thinks it has good data but it has a glitch and didn’t give good data.

That’s also why it’s still important to double-check your data and the sources you’re getting it from. ChatGPT is not perfect yet. Make no mistake. We’re in the first quarter of the artificial intelligence game. By the fourth quarter or by the end of the game, the things it’s able to do will be unrecognizable but now, it’s still early and it’s still not perfect. Like any new technology, you want to be careful to manage those risks and make sure you’re doing what you need to do to mitigate them.

Let’s get practical down to brass tacks in terms of managing those risks. Let’s assume it’s not going anywhere. It’s getting more sophisticated. It’s here to stay. I’m a construction company owner. I got a project right in front of me. I don’t have time to think about this ChatGPT stuff but I know it’s here. What are 2 or 3 practical steps I can take right away to begin to think about integrating this into my business to one degree or another?

It’s the easy way. Companies are introducing construction software and construction technology anyways through third parties. That’s already been happening before ChatGPT with Procore, DigiBuild, Autodesk and all those different things. The easiest way is to work with companies that specialize in software for construction and use artificial intelligence in their product because then you don’t have to worry about training your whole team on it. You have them use software or something like that. That’s the easy way. That’s what our customers do.

The back end is super powerful using all this crazy technology but the front end is easy and gives them these crazy results, cost savings, risk reduction and schedule improvements. One way is to use software and companies that specialize in this stuff but if you’re looking to individually or within your company introduce something like ChatGPT, find a project champion or a champion of some sort in your company who’s good at tech, who will go out and do some research on their own and who will identify, “On my projects, a good low-hanging fruit to start with is I upload my contract into ChatGPT and tell ChatGPT to act as a legal expert in construction in the state of Florida and break down this contract into the twenty most important bullet points.” That’s easy.

Disclaimer, your legal team should always be looking at this but it’s an easy way to see, “These are some of the things I can do with ChatGPT.” As a project manager, you’re trying to adhere to the contract and administer the contract. Instead of a 200-page contract, you have 50 bullet points. You could say, “This is what these sections mean.” Pick some small workflows and use cases, get a champion within your organization and have them start to build out and try these things on projects in a very low-risk way.

There’s no risk of having ChatGPT look at your payment terms and say, “What’s the best way to administer this?” That might be a bad example. Lastly, there are a lot of cool lunch-and-learn things you can do. Have a half hour where you bring in tamales or something for the team. We love doing this. For a half hour, we all look at ChatGPT and figure out use cases, how to use it and how to work with it. It ends up being a fun little thing that the team can do.

I like that as well. Nobody is saying here that ChatGPT is going to run your projects for you or anything like that. It’s an additional tool. If it’s used correctly and you learn how to use it, it can have a tremendous positive impact. Let me ask you one last thing. Is ChatGPT going to take over the world and kill us all?

ChatGPT will not. However, it’s important to see both sides in the things we’re going to do as far as curing cancer or even figuring out science and biology. Think about pattern matching. We will never figure it out. AI is going to change the world in that scenario but eventually, people fear AGI, which is Artificial General Intelligence where this thing is thinking as we do. Even going back to the agent thing I mentioned, if an AI’s specific goal is to make the world a better place, would they look at humans and say, “You are killing the world. I need to get rid of you all?” That’s the Terminator fear from the Terminator movies.

Not anytime soon, it’s not going to. ChatGPT is nowhere near effective enough or efficient enough. Sam Altman, the CEO of OpenAI, was on a podcast talking about that. Eventually, in 5, 7, 10 or 20 years, who knows where we will be? It is something to think about in being ethical and safe with new technology but that’s the thing with any technology. As we get bigger and scale the universe and society more, we have to be responsible. AI will be similar to that but we should be cognizant of both the good and the bad.

Robert, you’ve been very generous with your time. I appreciate it. Tell us a little bit more about DigiBuild and speak about how you are using artificial intelligence in the way that you’re running your business and helping your customers.

DigiBuild is a supply chain leader. We help companies in the supply chain to procure their building materials and anything supply chain-related. We will help them get better material pricing, keep their schedules on point and automate parts of the supply chain. We work with subcontractors, general contractors, real estate developers and anyone who’s procuring building materials. We can get them way better pricing because artificial intelligence reaches out to a much wider scale.

COGE 244 | ChatGPT In Construction
ChatGPT In Construction: DigiBuild is a supply chain leader. We help companies in the supply chain to procure their building materials, anything supply chain related.


We can tell everyone, “You might not be thinking about the lead time on this item from nine months out but artificial intelligence is. You need to order it soon because it’s a ten-month lead time.” We help to improve the budget schedules and reduce risks on these projects in a simple way. We use all this crazy powerful back-end technology but on the front end, it’s very simple. It’s easy to consume for construction managers. DigiBuild can help anyone who’s having problems with their supply chain or procuring building materials. We do that with a lot of companies.

How can people get in touch with you?

You can visit us at and request a demo. We will reach out or feel free to send an email to [email protected]. We will get on a demo call and show you how the magic works, the savings and the ROI that we give to our customers.

Here’s one last question. Is there any course or place that you would recommend someone goes to get some prompt training? Is there anything there?

Both Twitter and YouTube are great places to go. I don’t have specific ones off the top of my head. I’m happy to follow up with some for you if you want some links. Honestly, on YouTube and Twitter, there’s so much out there. Search ChatGPT prompt engineering. I told one of the biggest construction general contractors this. We were on a call. He was like, “Where does this lead as far as personnel? Where is this going to go?”

In many industries but especially construction, there’s going to be a new job type created called ChatGPT Prompt Engineer because if you can master ChatGPT, you can do things throughout the organization, “I can help you in estimating, course of construction stuff and even random code debugging.” You can copy and paste computer code into ChatGPT and it will tell you what’s wrong with the code or decipher the code for you and say, “This is what this does.” The point is there is going to be an employee position called ChatGPT Engineer. I was telling an ENR 400 general contractor that they should start preparing that because that will put them way ahead of the game.

If you can master ChatGPT, you can do things throughout the organization. Click To Tweet

I want to endorse that. You’re going to have a prompt engineer in business development, estimating, project management and your accounting site. You’re going to have prompt engineers all over the place in your business at some point here. I want to endorse that statement with a stamp. I stamped it. Robert, thank you for coming on. I love talking about this stuff. I do think this is legit stuff that’s going to change the industry. I want to encourage each one of my audiences to spend some time looking at both the benefits and the limitations of artificial intelligence like ChatGPT. It is going to be well worth your time. Robert, thank you for joining us here.

Thanks, Eric. It’s always a pleasure. I appreciate what you’re doing in getting the message for technology into the industry and being a leader in general in everything you do in construction. It’s always a pleasure to talk to you. Keep up the good work yourself.

Thank you for reading my discussion with Robert. I enjoyed that. I hope you did as well and you’ve got some takeaways as far as how you could begin to explore the world of ChatGPT if you haven’t already. Please, if you want to get in touch with Robert, go out to and learn more about his company. He is on the cutting edge of technology and how it’s impacting the construction industry for the good.

Thanks again for joining us. Give the show a rating or a review. One last thing. If you like episodes like this, hit me up at [email protected] and let me know because I would love to do more stuff on ChatGPT and AI if that’s something that you are interested in. Reach out to me. Let’s drive the content of the show so it brings the most value to you and your company. That’s what I’m all about. Let’s keep the conversation going.


Important Links


About Robert Salvador

COGE 244 | ChatGPT In ConstructionRobert is a project manager and construction technologist. After watching his father struggle as a small subcontractor in the construction industry, Robert eventually started his own construction company that grew into a nationwide commercial contractor.  Rob’s experience and passion for construction eventually lead him to search for a better way to manage construction projects using technology. He started DigiBuild in early 2018 to revolutionize construction projects, profits, and people.  He has managed over $500M in construction projects and was recently named to the list of “25 Rising Stars in Tech” By Crain’s Chicago Business. He is a frequent speaker and educator for groups like BuiltWorlds and the Construction Institute.