Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Empire Wind's Sudden Halt: What Happens When Megaprojects Get Paused?
53 Plays16 days ago
Join Jason and Carlos as they explore three key topics shaping the construction landscape:
⚡ Empire Wind Project Suspension: A look at New York's $5 billion offshore wind farm recently halted by a federal stop work order, and what project suspensions mean for construction teams.
🤖 AI Liability in Construction: Who's responsible when AI tools provide inaccurate information? The hosts examine the complex relationship between users, technology providers, and legal frameworks.
🏗️ The Long Tail of Construction: Should construction tech companies focus on the ENR 400 or the thousands of smaller firms? The hosts debate the merits of both approaches.
Follow The Offsite Podcast for weekly insights into construction and infrastructure.
00:00 - Intro
03:10 - Empire Wind Farm
15:15 - AI Liability in Construction
30:40 - The Long Tail of Construction
Check out the Off Site newsletter here
Follow Carlos on Linkedin | Follow Jason on Linkedin | Check out Aphex
Recommended
Transcript
AI Dependency in Workflows
00:00:00
Speaker
Once you get into a rhythm of using something and you start to ah rely on and trust the output, you're going to check less and less and less and eventually you're just gonna, you know it's better than you. So like I could do 10 things in the time I was doing one a few months ago. yeah So it's kind of like a lawyer whispering in your ear this is what you should do, but I'm not a lawyer.
00:00:21
Speaker
You're gonna still listen and ah take it and use it. So I can see why this is there's this horrible gray area of, it's still way better than what we were doing. And you know why they're saying this isn't advice, because you end up in a hole,
00:00:34
Speaker
contractor suing the vendor situation when they claim them for issues and then there's a like a if it's not advice it's more professional negligence before the to towards the person using the app to fast track their research or yeah whatever they're doing so it's a it's a real minefield of a of a situation but like there's no scenario in my head where you wouldn't use it you just need to be mindful so
00:01:04
Speaker
Welcome back to the Offsite Podcast.
Introduction and Empire Wind Project Overview
00:01:07
Speaker
I'm Jason Lansini, joined once again by co-host with most, Carlos Cavallo. ah Today we have three super cool topics, topical topics. We are going to first kick off with a look at the project that we profiled and then as a result got canceled. Well, not as a result. I'm claiming causation there.
00:01:27
Speaker
Not really true. The Empire Wind project that shortly after being broken down and profiled in our offsite newsletter ah got shot by Donald Trump. ah Next, we're going to have a conversation about who takes responsibility in the event of an AI software tool in construction making a mistake and revisit a discussion we had with ah Josh Levy on the topic.
00:01:54
Speaker
And then finally, we're going to debate a LinkedIn post that alleges that there is or that ah talks about the long tail of construction and the large um opportunity in smaller ah firms.
00:02:09
Speaker
Carlos, you're in London. I'm in Germany. This time next week, we'll be in Vietnam. Hey. Hey. It's probably the third time we've mentioned it, but the excitement's taking over. it's It's like the day before Disneyland. So yeah, two days. Yeah, we are yeah we're like a day or two from from heading out with our whole team.
00:02:30
Speaker
ah You packed your bag. Have you got enough tropical shirts? I've done nothing. I've done absolutely nothing. ah Yeah, I'm making a list. That's my strategy.
Travel Banter: Jason and Carlos' Vietnam Plans
00:02:39
Speaker
And then at the last minute, I pack the list.
00:02:42
Speaker
Yeah, i've I've been working really hard. I've been working really hard on my European winter tan. ah So i can yeah I couldn't be paler and whiter if I tried. attack I had a handy weekend in Barcelona last week. So you're hoping for some residuals on on that tan. I hope that tides you over. Yeah.
00:03:03
Speaker
When you put a tan on top of a tan, the whole tan lasts longer. It's my theory. So we'll see. Is it? Every time. Okay. i just I had about five Jersey Shore jokes go through my head, but I couldn't i couldn't make of them stack up.
00:03:17
Speaker
was a throwback. Nice. I don't know why that came to mind, but yeah, ah ah there's a lot of tanning going on. Paulie D. Nice. I wouldn't have been able to remember any of the names. That's all your reruns. You watched the last week. Yeah, I didn't do very well at university.
00:03:36
Speaker
Yeah, no, I was the same. Yeah, university, I was like, I could study or I could watch this. Yeah. Yeah, yeah, yeah, yeah, cool. That is a throwback.
00:03:47
Speaker
Can you watch it on Netflix or anything actually? ah With all that ton of free time I've got now, I might, I don't know what it was on. tv comedy Was it MTV or something? Was it, yeah, ah maybe it was MTV. It makes it sound old. People are thinking, that is there a channel for MTV? I don't know, I didn't. The reason I don't know is I watched it on LimeWire.
00:04:07
Speaker
They were like illegally downloaded. They're listening.
00:04:12
Speaker
ah Cool. ah i Well, diving in off the off the
Empire Wind Project Halted: Federal Intervention
00:04:17
Speaker
bat. So just one day after ah the offsite newsletter, famed publication generated from our team here at Apex, covered and profiled the Empire Wind Farm, ah it received a stop work order issued by the Trump administration.
00:04:34
Speaker
Are they reading it? um so to The industry must be so confused when they all read our newsletter. Yeah. quick So to let's start maybe with like what is this project for people that haven't heard of it and then we can talk about what actually happened.
00:04:51
Speaker
So by way of background, the Empire Wind Project is a massive offshore wind farm off the coast of New York. It's ah like 15 to 30 miles southeast of Long Island for those people that really know their geography in that part of the world.
00:05:07
Speaker
ah In its first phase, which was called Empire Wind One, the plan was to install 54 giant wind turbines. So um again, for people that really know wind turbines, these are 15 megawatt units from Vestas over, I think it's something like an 80,000 acre lease area ah for a total capacity of about 810 megawatts.
Empire Wind: Significance and Political Influence
00:05:29
Speaker
For those that don't know electrical engineering like me, that's about half a million homes that it can power. So a decent sized wind power project. There was also a second phase, Empire Wind 2, which is adding roughly the same capacity again, um ah maybe actually a little bit more.
00:05:47
Speaker
And it was the flagship project for New York's clean energy goals, which is something that it's pushing for. The project was approved ah um earlier this year um by some of the last things that were done by, I think, the prior administrations or administrators that were aligned with the prior administration.
00:06:09
Speaker
um It was billed of the largest energy infrastructure projects in New York. Yeah, in April 25, three to four weeks ago, ah the project was abruptly halted by a federal stop work order um while it was under construction.
00:06:25
Speaker
So thoughts, Carlos, a project that was like when we ah when we looked into it as part of the the research for the newsletter, we were particularly excited by what was a ah really cool project only to get it to watch it get shot, um which was...
00:06:42
Speaker
did you say um Did you see Trump's views on wind farms? ah Please tell me. yeah this one The first one is he said the noise for wind turbines give you cancer.
00:06:53
Speaker
Yeah, I believe that for sure. Oh, are you sure? Probably. Everything gives you cancer. So like that that seems fair. yeah Eventually it might do. Number two, it's bad for the environment.
00:07:06
Speaker
pretty not true the grand scheme of things. What part of the environment, like just the general aesthetic of the environment or? i I have no idea, but he's just said that everyone can go fracking as much as they want. So. but you can't see the fracking, so it's fine. So like, you yeah outside, out of mind.
00:07:22
Speaker
Yeah, that makes sense. yeah and then the third one he said when the wind stops blowing you won't be able to watch him on tv and that's an issue so that was his view on wind offshore wind projects um but no uh yeah massive project it's done in phases and there's a number of wind farms within the area it sounds like it is a temporary pause though um it's to do with uh like planning permissions and things like that. So very much sounds like it's delayed. Can we start as soon as it's tweaked slightly to become a nuclear power? Yeah. Once we just make those adjustments.
00:07:55
Speaker
Yeah, in four and a half years, when Trump leaves, it will start again. It's just been paused. So from an engineering point of view, apparently it's a very idealistic location because there are there's a decent level of wind and the the sea is relatively shallow.
00:08:09
Speaker
I think it ranges from 20 to 40 meters. And bearing in mind, you're not right on the coast. um yeah That's a bit of a benefit. It makes sense as a scheme. Obviously, New York is one of the big democratic hubs. So maybe, yeah, that's part of the reason. that yeah Yeah. So, yeah, the other project had like got full approvals. It's about a $5 billion dollars project investment this phase one. And they had um they had like work underway like Vestas.
00:08:39
Speaker
ah They had a $3 billion dollars debt financing facility that the development company had secured. And then, yeah, put a hold, which hold, cancel, pause, you know, whatever it might be.
00:08:50
Speaker
it got me thinking a lot about like what like putting the you know the team building its hat on. And probably more generally, you know we've seen around the world, these major infrastructure projects are often political footballs and they get you know started, then they get slowed down, then they get you know analyzed for cost and value, and then they get restarted
Impact of Project Pauses on Teams and Interests
00:09:15
Speaker
again.
00:09:15
Speaker
We've seen it in Australia with a bunch of the schemes like ah Suburban Rail Loop. In the UK, it's like the on again, off again HS2. So yeah, what is what's reality ah for teams that are on these projects when you know suspensions or cancellations or stop work orders happen? Yeah, there's ah I mean, there's a few different parts to jump into.
00:09:36
Speaker
Firstly, if you just think about the team who are ramping up to deliver a project, you've then got to probably redistribute people because if there's an indefinite time as to when this scheme may or may not sort of go ahead, you've got to offload the ah especially with the uncertainty of are you being paid?
00:09:54
Speaker
um It's also really hard to then not just like hold on to people, but hire people if you do need them, because it's quite unexciting to join a project that might not happen. Because but if you're an engineer on a project, you want to get really stuck in and and build up to that start date or package of works that you're managing.
00:10:11
Speaker
So that the team dynamic is quite difficult, but from a commercial point of view, there's obviously some pretty substantial sort of hotspots you can go down. um Firstly, like there's going to be some form of claim, right?
00:10:23
Speaker
And that can be anything from probably the most likely is let's make this cost of canceling so big that you don't cancel. So you'd come together and sort of say, wow, this 10 billion scheme is going to cost you four or if you don't do it. So like it's a stupid decision. So there's the...
00:10:39
Speaker
four Four plus you don't have the asset. As in it would be like the equivalent of 14. Yeah, exactly. It's burnt it's burnt cash. So there's there's one path which is make it too ah costly to cancel.
00:10:49
Speaker
And then there's the other side which is you could have pre you could have purchased your wind turbine. So there's going to be a... cancellation costs or they're probably going to charge you loads of money and still sell it to someone else.
00:11:01
Speaker
There will be some parties that i've been I've not been through this myself, but I hear loss of profit is a very hard thing to actually win, but at least a heavy disruption claim, the cost of staff that you can't redistribute, mobilization costs, pipeline, because organizations are only going to sort of secure jobs within what they can deliver. And suddenly they've got this big gap. So there's a an exercise that you would probably go down, this is to affect our business because it was a secured contract, for example. so What have you seen typically happen around like there's this moment, I presume there's some obligations to mit ah mitigate costs and minimize costs when you receive those sorts of things.
00:11:39
Speaker
In your experience, have you seen like the client that might issue such a notice be quite prescriptive about what should happen with people? or do they be but Or projects and teams left in this like open-ended spot of like, does that mean we should get rid of people? Or are we waiting for this to restart? Do you end up in this kind of limbo spot of we don't really know what to do? Yeah.
00:12:03
Speaker
Obviously, there can be... You could have detailed termination clauses in a contract that might actually sort of state everything in black and white, which is... But usually, like, yeah, just, it usually is like a, there's like a pause or a stop before a term, like it's never like a zero to term, well, it doesn't, and ah'm not that I've seen a zero to termination, like, moment. Yeah.
00:12:26
Speaker
I think you'd be in a ah bit of an awkward gray area because if you're going down this path, and let's just say it's kind of a slow stop to a hard stop rather than like an immediate stop, quantity surveyors or commercial leads are naturally going to try to create as big a claim as possible. Mm-hmm.
00:12:41
Speaker
And then you're probably not going to just cut everything immediately because the claim is going to be built on this cost that you're going to keep expending over time whilst the situation is resolved.
00:12:52
Speaker
But all of that cost is also at risk. So if you then lose the battle at the end, yeah you had subcontractors and then suddenly you're losing everything. It's it's a real gamble.
00:13:02
Speaker
But if you do get rid of everyone, the claim gets small. So it's yeah it's a bit of ah a gray area in the middle. Yeah, you have your whole team. And yeah, when there's like that lack of clarity, they have the whole team going, what do we do? Where are we going? And then, like you said, you've got a whole supply chain that are asking the same question.
00:13:18
Speaker
And you'd probably play off the uncertainty by asking for instructions to maintain a crew or maintain something to give you some kind of ah guarantee is a strong word, but likely cost recovery because you're still getting instructions. But yeah, you wouldn't want to be completely at risk because...
00:13:36
Speaker
Yeah, some of these schemes, you run out of cash, are you really going to arbitration? Yeah, you'd you'd almost play the pretty hard line of there's the contractor of like, well, we'll just stop everything until we, yeah until unless you tell us otherwise sort of thing.
00:13:50
Speaker
And we'll let the supply chain go. And we'll let like basically putting the pressure back on the client to be like, if you don't want us to do that, you have to tell us not to. Yeah, exactly. And if you think some of these schemes are so massive that the knock-on effects can be contract just share price starting to drop and because of loss of revenue. And there's so many factors at play around one of these major schemes cancelling.
00:14:12
Speaker
But there's also this there's there's also a whole safety consideration, like the contract is still responsible for the site. So you could have, you know, you could have temporary works in a state that might not be like best to leave there. You know, there's all that stuff you have to kind of do an analysis of.
00:14:27
Speaker
Yeah, there's a lot of work that comes off the back of it. Yeah, I've not been ah in an actual termination contract before. um We've been in like slow down or we need to reduce what we're doing over a period of time, um which is kind of a similar game where you're constantly saying, well, we still have to function. So most support departments, they're there, whether you're doing 10 million or 100 million um in a period of time. but um Yeah, and when when they're politically driven, like this may or may not be, judging by the three reasons for cancelling it that you just laid out, I'm guessing yes.
00:14:58
Speaker
When they're politically driven, it does feel like they could just be turned back on like the drop of a hat. and And so you have to, yeah, I think the contractor's got to play that game of like, well, we'll just pack up unless you tell us otherwise, which puts the pressure back up the chain of like, if you do plan to turn this back on and you don't want this to cost an absolute boatload where...
00:15:17
Speaker
Because yeah, are you as the contractor thinking like if you then wanted this turn back on, it'll then have all these costs? But if you overplay that hand, they could retender the whole thing. thing Like you you could you could be in this spot where, you to be really careful of your of what of what cards you're playing. Yeah, if you go too hard on a what is seen as a pause, you might not be coming back at all.
00:15:38
Speaker
ah Yeah, completely. yeah it's so It's one of those situations where there's going to be lots of kind of honest conversations, but then the formal communication is very sort of- It's vague. Yeah, vague as hell. Yeah.
00:15:49
Speaker
Because no one wants to cap themselves later. Do the things as per the contract. i feel Yeah, exactly. And then in the background, meeting it's like, this ain't happening, boss. like Get rid of everyone. yeah Literally.
00:16:02
Speaker
So relationships are crucial at that point, I'd imagine, between individuals. Yeah. And so if anyone else wants their project canceled, let us know. We'll do a profile in the offsite and and surely within a day or two, it'll usually get they usually get killed.
AI in Construction: Liability and Errors
00:16:20
Speaker
Cool. So moving on ah with this like imminent and increasing growing wave of AI functionality and purpose built AI tools and platforms ah looming and coming into the construction industry, the number of places where it plays a part in how projects are delivered, decisions made on the projects ah growing.
00:16:44
Speaker
And at some point, you're going to inevitably have this negative impact by someone, by giving some false results, hallucinations, someone following something that they might have needed professional advice on.
00:16:57
Speaker
ah And this brings up a question of who's liable in that situation when it arises and and how the AI software industry is handling this and then maybe what the broader software ah industry does in this regard.
00:17:17
Speaker
We asked a very similar question last year. we had Josh Levy from the CEO Document Crunch join us for a conversation. We asked him, I think, quite a similar question. And ah Oli will cue that clip up now.
00:17:33
Speaker
How do you square or how do you answer the question of ah legal liability regarding kind of perceived advice with the kind of giving advice or making decisions? And as you go further and further down that path,
00:17:51
Speaker
Yeah, it's a really great question, Jason. So like let me ask you a question here. um how do I'll ask a rhetorical question. How do these construction companies deal with giving legal advice or making recommendations when these folks in the field are engaging with this unsupervised everyday anyway?
00:18:10
Speaker
Yeah, it's me being a... boy go We used to call ourselves Bush lawyers. ah But when you get into this... Bush lawyers. I love it. I've never heard that before. But like look, here's the deal.
00:18:21
Speaker
We're not trying to give advice. We're not trying to replace the need for legal, but we are trying to fill ah yeah that does exist, which is to make some of this easier and easier to understand. So we we are not trying to become lawyer.
00:18:37
Speaker
I would say that what we're really trying to become is like a supercharged quantity surveyor or project administration assistant or something along those lines. But of course, as you know,
00:18:48
Speaker
Like there's folks in the field that are making these own recommendations every day with zero qualifications to do so, right? So this is just to supplement decision-making, but it's not to ah replace the need for appropriate counsel. And in fact, guy making this easier to understand what we're actually seeing, what our customers will tell you is actually teams are more aware of when they should engage to get the appropriate advice because this becomes like something that's like not something that we should just render a decision on. So ah definitely not trying to replace the legal side.
00:19:26
Speaker
is really So let's think about the, if you were to bucket the the response or defense or the position from from Josh there, which I think is broadly reflective of what I've seen in the marketplace of other ah tools that are either similar, competitive, or in different areas of the construction value chain to to Document Crunch.
00:19:48
Speaker
is there's one argument which is this is happening in the field right now from people that are not lawyers, you know engineers, project managers, construction managers, reading documents, making decisions, and wouldn't it be better to have some kind of advice or support or whatever?
00:20:04
Speaker
And then the second is that I've seen others make, Josh didn't directly make it there, which is like, well, we're giving this advice to the professional who then you know actually decides what they do. So the the person is holding that kind of responsibility.
00:20:19
Speaker
Are they valid? I think um just to like give it a bit more context for anyone listening and a kind of what we're talking about when we say ah using AI in this kind in this sort of sense. When anyone starts on a job, you have your works information. You've got stacks and stacks of documents that build up like the scope of the project.
00:20:38
Speaker
Most people would sort of Their first week or two is probably reading through and like you said the other day, Jase, you would tag the sections relevant to your package so you've got that sort of go-to set of information.
00:20:50
Speaker
And then let's say you're procuring a package, it's trying to identify potential risks, um you want to look at liquidated damages, you would zip to the sections you know and then you do a quick control F to try find bits you might have missed previously. That's the process because there's way too much information to read through it every time that you need to go through one of these exercises.
00:21:09
Speaker
yeah With something like Document Crunch, you can upload all that information and it's a very intelligent way of pulling everything that might be relevant to a certain subject. And it was gonna a million times faster than you and probably more accurately.
00:21:22
Speaker
Back to your question, the reality is, When you're using a tool like that, we use tools like that for various...
AI vs. Legal Advice: Navigating Limitations
00:21:31
Speaker
90% of what I do is AI. You're not even Jason anymore. yeah Your name's Claude.
00:21:36
Speaker
um So you you once you get into a rhythm of using something and you start to ah rely on and trust the output, you're going to check less and less and less. And eventually you're just going to, you know it's better than you. So like I could do 10 things in the time I was doing one a few months ago. yeah So it's kind of like a lawyer whispering in your ear, this is what you should do, but I'm not a lawyer.
00:22:01
Speaker
You're going to still listen and ah take it and use it. So I can see why is this is there's this horrible gray area of, it's still way better than what we were doing. And you know why they're saying this isn't advice? Because you end up in a whole...
00:22:13
Speaker
um contractor suing the vendor situation when they claim them for issues and then there's a like a if it's not advice it's more professional negligence before the to towards the person using the app to fast track their research or yeah whatever they're doing so it's a it's a real minefield of a of a situation but like there's no scenario in my head where you wouldn't use it you just need to be mindful Yeah, I think the problem, i think that, yeah, you would definitely use it. I think the the risk is actually just with the company when something eventually inevitably goes goes wrong here.
00:22:47
Speaker
Like, you're right, like just to to echo back what you were saying, There are different workflows or things that you would do with AI. And so you have this like, this kind of like, you can think of it as like enhanced search.
00:23:01
Speaker
ah Behind the scenes, what happens is that there is this ah um tooling or ability to do what's called retrieval augmented generation. So basically there's like this way of chunking up all the data and making it more semantically searchable.
00:23:16
Speaker
So it's like, ah Really simply like control F on steroids is what you can kind of think of that as. And that's just helping you find stuff.
00:23:26
Speaker
And there's probably relatively low risk with that. Next game of it, but next level of that is probably something like, and I'm oversimplifying this, but like summarization. So like read a bunch of sections from all these documents and just give me the kind of like two paragraph summary of the thing.
00:23:42
Speaker
And there you've got risk of like stuff not appearing that maybe was relevant that you're like missing. And then you've also got the like, ah if it's not properly on Rails, it might introduce stuff that wasn't in like the actual content.
00:23:56
Speaker
Equally risky is like the person just reads the document and misses the section. So it's like how much more risky is it than the person who's not actually read the document. But to your point, once you start using this and you go, oh crap, that's actually super useful. That saved me a bunch of time.
00:24:13
Speaker
You just keep kind of like pushing how far you could like take it, how far it can take you. And as the trust grows. Yeah, that's right. And you're a hop, skip and a jump away from after you ask it to summarize a bunch of obligations around ah RFIs in a document, you well, how do I submit it?
00:24:31
Speaker
ah Or then you go, I've just had this issue on site, rate write the ah RFI. Like you you just, you really just start like putting more and more of the like making sure we're compliant with the whole thing and the whole process off of your hands and into it. And you could see how that, or you can see how suddenly someone that has been doing this for six months or a year and is like really confident in some of those base underlying capabilities, then you know writes and says, I've just had a subcontractor who had done this.
00:25:02
Speaker
what should What should I do next? You know, and that yeah now we're like- Is this person liable for X? No. Cool. Not gonna think about it. Correct, yeah. As you can see easily how you like cross that line.
00:25:13
Speaker
And we don't, today, like i review contracts with AI internally. And when you're forced with that decision as the user, you're like, I could get the answer in seconds or I could find a lawyer. Like, on this is ah very rarely like a lawyer just sat around like, got any questions, I'm here for you. To find someone in a head office that knows who the lawyer is, that you have to ask the question, they're usually total pain in the backside.
00:25:40
Speaker
I'm very busy, I'm very important, I'll get this to i'll get back to you in four months. So you're going to you're going to ask the question and you're going to action off the back of the question. So i think um i think Josh is totally right.
00:25:54
Speaker
I think it is way better than the person in the trailer or on site not doing any of it now. But as they get more comfortable, they're going to rely more on it and they're going to it's going to cross into that space where they would have typically asked legal advice and they're going to rely on AI to do that. And you can also imagine a situation where it if it becomes very normal to use it and it does make a mistake, everyone will say it's not their fault and blame the tool.
00:26:18
Speaker
And then you're in a really weird spot then where no one feels responsible or is held to that decision that they made using this tool as a supplementary sort of research buddy or whatever it is. Well, with your background, you'd you'd you'd know exactly you'd see this exactly like when con when something goes wrong on a project, what's the contractor doing? Who can I possibly fling this mud on yeah that isn't me?
00:26:44
Speaker
Designers, the tools that we use, that does happen. that does happen There isn't a strong legal framework around this from what I can tell. I don't know if you found anything different. No, it'll be really interesting to see like the first construction specific case go to court. Not that we want to see that go to court, but yeah there's no like, don't know, precedent on liability um and how that will really sit. We're all making assumptions. is ah One thing that I'd love to know from a lawyer is just by saying this is not advice, is there is that enough?
00:27:16
Speaker
Yeah, no, yeah. Because you could just make that statement with everything and it's never your fault. That's like the old like off the record statement. You say that and then it's like suddenly I don't hear anything you're saying. Like an e email with PNC at the top and then someone forwards it. Is it really?
00:27:32
Speaker
It'll be an interesting one to follow. Yeah, the because ah there are like there are kind of these loosely defined terms in a lot of contracts right now, you know even as a software vendor around ah around risk indemnity.
00:27:45
Speaker
Those clauses can be very loose and there is an established precedent that is tested through case law. ah you know what yeah Where does the risk effectively sit?
00:27:58
Speaker
um And so... presumably you've got to imagine what happens here is that yeah some mistake does eventually get made in the world and this gets tested.
00:28:11
Speaker
um and then And then the marketplace around contracts and everything start to to shape off off the back of it. the The thing is, like if if ah if an employee makes a mistake, what happens?
00:28:25
Speaker
you You work through that mistake. i yeah Let's just say you're the ah professional indemnity insurer. Would you rather AI or junior quantity surveyor?
00:28:37
Speaker
like They might see that as a less risky thing than the staff that are always missing things and hi missing deadlines and doing things wrong. It might actually be seen as the less risky thing to do, but everyone's uncomfortable with the fact that it's AI making a mistake and not a human. Yeah, probably starting with the contractor's ah insurer presumably we're talking about. yeah And they probably need a ton of actuarial data in order to make that assessment. And so in the interim, when you don't have any of that data, you're in this spot where, you know, the way that typically happens is that insurer then looks at like what other insurers could actually split this blame.
00:29:13
Speaker
So because, yeah, the claim will come for the contractor through some issue and the the contractor will have, you know, let's say there was, I don't know, quality issues off the back of some advice or whatever that resulted in rework cost.
00:29:25
Speaker
and the contractor decides they're trying try claim under their construction insurance policy that insurer then is possibly going look at that incident and then try to like do you think that's how this gets tested like the insurer then tries to shoot sue the the advice giver yeah it's a reasonable path because they as soon as they look into that situation and realize ai was used to inform the decision that's going to be something that you sort of grab and run with you'd expect Because that there they're insuring against usually some degree of negligence or whatever at the individual level on the project.
00:29:59
Speaker
Yeah. So that they they the and the insurance company that's coming for you would grab it. For your own insurer, they're probably feeling a bit more comfortable knowing you use it to reduce the likelihood of gaps and mistakes. Once they get the data.
00:30:13
Speaker
Yeah, yeah. Once there's enough runs on the board, because it's the same. i listen There was an interview recently with Warren Buffett at the um at the annual shareholders meeting talking about um the insurance risk of ah ah for vehicles with the increase of autonomy.
00:30:30
Speaker
And presumably you end up in the situation where ah there's a lot more one a lot less one-off incidents because you've got you know, ai doing something better than the crazy mix of people.
00:30:42
Speaker
But you end up with this like tail risk of like one thing that's super expensive because it impacts lots of people as well. the whole The whole risk profile of the insurance policy changes pretty significantly.
00:30:54
Speaker
Here's an idea. On a project, they should just run a little test with each individual. And at the end of that test, they go, you need AI. So you will use AI. Oh, you're fine. Yeah, you don't need it.
00:31:07
Speaker
Yeah. yes well We've all got that person on the project. with Don't ask Bob anything. but but but bo plus a i bob class a i bob you Bob plus equals a Steve.
00:31:23
Speaker
what you don't want to be is the like just gym plus three a eyes that gets you up so bob We'll give them ChatGPT Pro. Yeah. All right, let's move on. But I think there's there's a lot to come on that topic, which will play out probably over years, I would say.
00:31:44
Speaker
um So jumping to LinkedIn, ah your favorite place, um there was a post that we ah we discussed recently internally from a post by Martin Roth. So for those that don't know, Martin but was one of the early go-to-market team members for for Levelset and then became the chief revenue officer.
00:32:08
Speaker
Levelset was a ah US-based construction technology company that did ah lean payment management. They grew to pretty sizable revenue and were acquired for something like 500 million by Procore in the last handful of years.
00:32:25
Speaker
So Martin knows a thing or two about the construction industry and posted recently that effectively construction ah is the long tail.
00:32:37
Speaker
And what he meant by that is like most construction technology companies chase ah the ENR 400. So for anyone outside of the US that's like the top largest 400 contractors. You might call them as like the tier one, tier two in other regions um there.
00:32:56
Speaker
They are notoriously hard to sell to. They take a long time. They take they run endless pilots which you know of tools and and technology. But outside of that are hundreds of thousands small medium-sized businesses that not enough people focus on.
00:33:14
Speaker
And the vast majority of construction companies have less than 50 million in revenue. Yeah, it basically drives a, well, there's like a ah big focus on the biggest contractors, which may be misplaced.
00:33:26
Speaker
And it probably drives a lot around how construction technology companies should think about their go-to-market, think about the market that they're actually selling into. and actually drives ah thinking about a whole bunch of topics. So um starting there, well, maybe start with like testing the facts.
00:33:45
Speaker
Do you agree that that's like true? Because I can see cases for and against. Yeah. If we take it in a ah black and white sense, I can see why the idea that like there's more money to made in like 10,000 companies doing $10,000 of revenue to like going after 10 customers doing much bigger pool that because you've got the mass market there. I think when I really thought through this, I could see the argument, but I think there's a couple of factors such as how complex the tool is or the piece of technology. And then ah off of that, you've got like support and, and everything that comes with it.
00:34:21
Speaker
And also if you, okay, if, but if I just go back to the actual, uh, the thought rather than trying to sort of frame at the end, The benefit of going after the bigger contractors for me is, yes, you're going to have these annoyingly long procurement cycles.
00:34:36
Speaker
You're probably going to have to go through the pilot process. you're going to have to like do the dance, jump through hurdles to even get like your first invoice being generated. yeahp um And then you need you need cash. You need all sorts of things to be able to make that happen in reality.
00:34:51
Speaker
But... You're then in a kind of, you're not just ah a vendor, you're more of a partner. You're building and deepening a relationship with a larger organization. You're using that logo that everyone knows to to jump to the next one and supporting a big project rather than a thousand small projects.
00:35:12
Speaker
with a complicated tool in inverted brackets is a much more manageable job with a normal team size, rather trying to support thousands and thousands of projects. If I flip the argument, I think the benefit for the small projects is there's almost no red tape um you're probably in the door pretty quickly if you speak to the right person. They're probably super loyal because they probably don't use that much tech. And if you have an actual customer facing team who's on the end of a phone call, you're probably in like ah a pretty strong position, assuming that the tech's working for them.
00:35:44
Speaker
But its that crowd of people, how you actually market to those, Because like if we looked at LinkedIn, it's there's loads of people on major projects with big customers that want their nice professional corporate kind of look to it.
00:36:01
Speaker
How you actually get to a market like construction with all these small builders to make enough generate revenue quickly um is the the question mark for me that I don't know how you would go about tackling, which wouldn't take years and years and years and years and years. Yeah, obviously that's the part of the market. We're actually doing the thing that he said not to do.
00:36:20
Speaker
to be completely. Yeah, yeah, yeah. but like So yeah, maybe to to add some more data to this. So first of all, like i'm ah I'm probably like a big Martin Roth fanboy. I think a lot of the things he says are spot on. And I think he's technically correct on all these points as well.
00:36:38
Speaker
ah If you just look at the US and Europe, how do you measure the size of the construction industry? Well, there's like at a high level, you've got the the number of companies, the number of firms, the amount of people they employ, and then maybe the construction volume that they perform.
00:36:52
Speaker
And by all of those measures, it is the long tail. If you took the small, medium, so anything less than 500 1,000 people, It's like 99% of companies in the construction industry are them.
00:37:04
Speaker
Something like 80 to 90% of the workforce are them. And roughly 80% of total construction output volume are them. Factually correct that it is the long tail.
00:37:16
Speaker
And that's where like most of the people are. That's where most of the construction volume is and basically all of the firms. But... The bit where I i think is i kind of missed in the in that ah statement is that the volume doesn't necessarily match with where the control is and where autonomy is. And what I mean by that is If we think about how any major project is delivered or what are all these SME companies working on?
00:37:47
Speaker
Well, we know that the way that construction projects get delivered is major contractor gets something, they sub out a part of it, that person subs out a part of it, that person subs out a part of it, et cetera, et cetera, et cetera.
00:37:57
Speaker
So you end up in this spot where you're weirdly triple counting, quadruple counting construction volume as revenue for all of these companies in the stack. Because of that, you're going to get heavily weighted, but these people,
00:38:09
Speaker
not all of them, but in ah if you were to redistribute the percentages based on how many of them are actually working on or contributing to projects that are controlled by a major contractor, the numbers change a lot.
00:38:23
Speaker
And oftentimes when you're selling software or deploying tools into a major contractor, you are capturing all the people on that project. Some of them you are. some you know If you're making design tool, different workflows or different tools have different ah users and workflows and audiences.
00:38:39
Speaker
But at least in the space that we play, like you're goingnna you're deploying something with a major contractor, you're on that project going capture some large percentage of their supply chain as the users as well. Whereas if you were selling something into the supply chain, if you if it's something for them to do in their back office that no one else you know has to care about brow or interface is fine. But if it's how they deliver their work, well, suddenly you might be bumping up against the fact that they don't control some of those ways of working on a project.
00:39:07
Speaker
It is definitely the long tail, um but there is, i think, an outsized impact or control asserted by the smaller number of large contractors on that are delivering these large projects that are controlling how technology and ways of working happen on the project that then are impacting the sub, the sub of the sub, the sub of the sub of the sub.
00:39:31
Speaker
and so I think the answer is probably dependent on you know what piece of technology you you're building, what workflow that kind of captures. Because yeah, you as you said, if you can sell to the smaller company, you don't have to go through five pilots and do 17,000 business cases and and all that.
00:39:51
Speaker
um But if they need it to interface or be used on that project in a way that touches the main contractor or a higher level authority, well, you're you're basically back in the same spot. Yeah.
00:40:04
Speaker
It's interesting point with um the doubling and tripling of of value because of the main contracts. And obviously, that's partially why contractors' margin is so low. Most of the work isn't actually them delivering anything. It's just correct numbers on their books.
00:40:16
Speaker
But I'd still say that you would still, in order to get that, youre yeah you're still going for the main contractor no matter what in that sort of instance. But if you actually, let's say we isolated small projects or small contractors on contracts that they're not working under one of these big might make mega schemes, I think that'd be really hard graft to go and sell five licenses here, 10 there.
00:40:40
Speaker
ah to make any real traction. It depends. It depends on what the piece of software is. like how How complicated the tool is. like ah i'm Not even complicated. it's just like what the workflow is. Like if you take level set as an example, which was like how people get paid or chase liens, that's a problem that a subcontractor or a sub of a subcontractor yeah definitely going to have and that they need to solve for themselves.
00:41:03
Speaker
So that is definitely deployable. and inside themselves. Yeah, but if you think about more like site tools. Yeah, site tools or anything collaborative, systems of collaboration, big big time. you You're in a spot where you kind of need to go with the controllers.
00:41:17
Speaker
It's a really interesting point. Recommend people follow Martin. you going for dinner with him soon? No.
00:41:24
Speaker
Good stuff. Read us out, mate. Right. Thank you very much, everyone, for tuning into today's show. If you did enjoy today's episode, please do think about liking this video or following us on your chosen podcast platform.
00:41:39
Speaker
We really appreciate your support and we'll see you all next week. Bye-bye.