Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Southeast Asia's Construction Boom: Giants You've Never Heard Of
47 Plays11 days ago
Join Jason, Carlos, and special guest Sam Hanisch as they dive into three developments shaping the global construction landscape:
🏗️ Southeast Asia's Construction Giants: Explore the $500 billion Southeast Asian construction market growing at 6% annually. Discover major players from state-owned Indonesian contractors to international specialists, plus massive projects like the $40 billion Trans-Sumatra Highway.
🤖 Trunk Tools' $40M AI Revolution: Analyse the construction AI startup's massive Series B funding and their RAG platform connecting construction documents, drawings, and schedules. Examine what this means for AI in construction versus existing solutions.
⚡ The Future of Planning Professionals: Discuss a controversial post claiming expensive planners earning $200-400K are becoming obsolete. Evaluate whether technology will replace traditional planning roles or transform them.
Key Timestamps:
00:00 - Introduction
02:26 - Southeast Asian construction market analysis
12:40 - Trunk Tools funding and technology breakdown
34:29 - The obsolete planner debate
Check out the Off Site newsletter here
Follow Carlos on Linkedin | Follow Jason on Linkedin | Check out Aphex
Recommended
Transcript
Introduction and AI Tools Discussion
00:00:00
Speaker
i was gonna say would you get it would you buy it on your project you're out on the ground would you give it a go oh well we talked before this offline before we started recording about how much money i've spent on every ai tool so 100 you're just in your office a i tools yeah one of these on day one immediately just to search and query and and then extract information um yeah yeah i don't know you can function without that now
00:00:29
Speaker
Welcome back to the Offsite Podcast.
Podcast Introduction and Guest Appearance
00:00:31
Speaker
I am Jason Lansini, joined once again by co-host Carlos Cavallo and special guest Construction Sam from Fresh Out of fresh out of ah of Bali.
00:00:44
Speaker
This was a lesser known TV series. So what Bob the Builder constructs today. Today, we are talking about a couple of super interesting topics, breaking down the biggest contractors working in
Episode Topics Overview
00:00:57
Speaker
Southeast Asia.
00:00:57
Speaker
Next, we're going to dive into a massive raise of 40 million usd Series B raised by construction technology AI company Trunk Tools. And then finally, we're going to discuss a fairly controversial post we saw online claiming that planners are going to become obsolete in the near future.
00:01:18
Speaker
G'day Carlos and Sam, thank you very much for joining us. Yeah, what are your thoughts on, are you a fan of the construction, Sam? There's going to be an ai picture made almost ah immediately after this episode.
00:01:33
Speaker
And ah knowing Jason, it'll have Bali reference or something and a hard hat. Just relaxing. It'll be you on a cabana chair, yeah relaxing, getting laundry delivered.
Southeast Asian Construction Market Overview
00:01:47
Speaker
That is what he does all day, every day, right? I do feel like someone comes with my laundry every time we get on a call, Carlos. Yeah.
00:01:56
Speaker
Yeah. For context, anyone listening, Sam used to live in Australia. He now lives in Bali. think they might've got that from Bali. Just in case it wasn't clear. I think it's when, you know, when people say living the dream, I think they've, that's where it's, that's the origin of, you know, you think like origin of the phrase.
00:02:12
Speaker
Yeah. That's the origin of the phrase. That's pretty good. Nice. Well, thank you very much for taking the time on your, ah where you could be doing so many other things in your evening. um the The choices are endless. Thanks for joining for the combo. I guess the, ah for those who don't know, the Southeast Asian construction market ah is, I think slept on a lot. It's roughly about 500 billion a year and growing at about 6% a year.
00:02:39
Speaker
ah yeah I think most of the conversation that we have had and we see probably happening ah online or in the industry focuses on US or North American, European markets and ah the Middle East.
00:02:53
Speaker
The makeup of the market, given it's a collection of lots of different countries, is ah dominated by state-owned contractors like those from Indonesia and then a mix of private companies from countries like Malaysia or Thailand. Given ah it's an area that doesn't get discussed enough, I think most people are totally across who the major players are, what types of projects are happening there at the moment. And given Sam, you're positively a local, thought it'd be interesting to yeah discuss some of those contractors and interesting projects that are happening in the region. We shared a top ah contractors list. And like I said, there's about 500 billion worth of construction volume. About 200 of that is inside infrastructure a category of work.
00:03:39
Speaker
And there's some really massive projects like the Trans Sumatra Highway that's about 35 to 40 billion US dollar equivalent. And the size of some of these contractors rival the largest contractors or large contractors in in other markets. So yeah, and diving into that, Carlos, did you have any things that you learned? Yeah, so generally, I mean, you've already mentioned the size of the market. It's quite impressive how much work is going on and it's relatively under radar.
00:04:07
Speaker
Obviously, we're focused on use outlets and things like that, which you're focus more on. within region projects, but it's pretty substantial from an in- five start infrastructure perspective. There's a ton of data centers in that part of the industry, which is growing rapidly. But it's quite interesting, the split. I was quite surprised. I think 60% or something like that of the spend is actually in in Indonesia alone.
00:04:29
Speaker
yep There's a huge amount of work, projects going on. It feels like if you wish strategically placed Sam in Bali for that. um but i had and I had no idea, if I'm being honest, ah the the volume of work going on there.
00:04:41
Speaker
Yeah. I mean, Indonesia, i think, is in now the fourth largest economy in the world. So it's sort of not surprising, based on my observations of the infrastructure here, that there's ah some money getting spent in that
International vs Local Contractors in Southeast Asia
00:04:53
Speaker
space.
00:04:53
Speaker
Fourth largest economy in the world. Sorry, fact check, fact check. Go for The population always catches people. I think it's like a 300. It's approaching 300 million in terms of population. It's a massive population center. Yeah, 300 million. um But yeah, no, Sam, to your point, like I think people do do underestimate the size of the, at least the the region and country by by population.
00:05:15
Speaker
Must be a, yeah, 16. good fact check. Maybe it was a projected in the future or a population. Um, But yeah, i i don't know if their account really is a local or an observer to these contractors. and of None of these are getting into the villa or the swing game over here in Bali. but Swing game?
00:05:37
Speaker
Yeah, the swing. know the swings? Yeah, I know about the swings. I'm probably not blown away as it sort of reinforced how many were state-owned. It feels like a lot of the companies like doing these big projects are like X state-owned or yeah some involvement there. Yeah.
00:05:54
Speaker
Also, the other thing that blew me away was the size of the contracts because we looked at a local list and then an international list. um It felt like the same mix of international players, but the contract values really dropped off for the local players. Do you think it's a capability thing or is it...
00:06:09
Speaker
the the clients that are dealing with those large contractors. Are you saying that the looking at the construction volume by a contractor ah favored international contractors versus local?
00:06:21
Speaker
Yeah, looked that way to me. Which, you know, in some ways is not different to what we kind of see in other markets. Like if you look at Australia, I think if we'd done that breakdown, you probably get to similar results. Yeah.
00:06:35
Speaker
I think it was nice to see the the actual mix. So there are quite a few local contractors doing very well, which is good. There are the typical contractors, ah European and American contractors in the mix.
00:06:48
Speaker
But if you look at it, they are kind of the specialists at the head of their game in each part. So you've got Vinci doing airports like that's consistent around the world. a capability piece. Yeah, we've got Excite doing data centers.
00:07:00
Speaker
yeah We've got Floor doing like pharmaceuticals and and oil and gas. We've got Boigs doing tunneling. We've got Hockchief doing tunneling. So it does feel like the major players in the game are in there doing the thing that they do best.
00:07:14
Speaker
But it's good to see that you haven't got Western contractors doing not their typical projects, which means they're just winning everything um and beating out the local competition. So it feels like it's kind quite a reasonable placement of these organizations.
00:07:28
Speaker
Yeah, there's definitely some sizable local contractors in the region. Like obviously, Gamuta is a very, very large area in the space. I know a little bit about it and we've we've we've seen them make large investments into the Australian market. Sam, you probably know even more about about the Gamuta business. Yeah, I think ah it's probably the most recognizable for any of the Australian construction co-war with Gamut has probably been five years, five plus years in Australia now.
00:07:57
Speaker
um And of the list, I think they ah they're the only Southeast Asian contractor working in the Australian market. So sort of stands out why they're their um balance of work is ah significantly larger than the second highest.
00:08:12
Speaker
Interestingly, talking about specialists, they really stand out as the kind of metro and like rail transport gurus from the region. And even coming, they can bring that technology over into Australia. Plus tunneling as well, I think. I think they they're at the forefront of tunneling tech. I believe they were the inventors of some like the first AI powered TBM.
00:08:32
Speaker
Yeah,
Major Players and Projects in Southeast Asia
00:08:33
Speaker
okay. the um Yeah, definitely they... they probably carve out like a special position, you know, which enables them to make the investments they've made to enter like the Australian market.
00:08:43
Speaker
One that jumped out to me, which I don't know if anyone did a deep dive on, was ah like second highest on the list is a company called the Italian Tide Development ah Company.
00:08:53
Speaker
Yeah, the name stands out, doesn't it? Yeah, I immediately it was like, hmm. It's like fusion food. It's like the worst restaurant you've ever been
00:09:05
Speaker
to. So I did a bit of a deep dive. It founded in the 1950s between Thai doctor and a Giorgio Berlageri.
00:09:17
Speaker
And um they actually built. ah Had a few pints and thought, oh, why not? No, they built the the main airport in Thailand in um in Bangkok. Oh, wow. They've done a ah ton of massive projects, but the company I believe is in like financial crisis right now.
00:09:33
Speaker
so That's what I read. the It seems like one of those companies said, oh, they built the airport. I'm like, cool. I think that happened quite a while ago. If that's still the most notable project, you're probably not in the best spot.
00:09:44
Speaker
They currently have total debt of about $3 billion US equivalent. um And are losing ah about 170 million u s equivalent per year.
00:09:57
Speaker
And they've got ah they've got a pretty epic project that's crushing this's like really dragging them down on a ah port project. 20,000 employees um that are allegedly awaiting salary payments.
00:10:12
Speaker
Yeah, I'm pretty sure, I'm trying to fact check myself right now, that they had a skyscraper that they built which collapsed as well. So you can imagine the knock-on effect come commercially from that. I had that detailed somewhere. So not even, so not a good restaurant and and maybe not also the best restaurant.
00:10:31
Speaker
were they number two on the list? I they i think they were number two on the list. I'll do it. bit of a gap Maybe that answers your initial question, Carlos, about why the international contractors have got so much work.
00:10:44
Speaker
Yeah. In better news, did you see that Penang light rail project that's going in that's just starting with Gamuto? No. I think it's probably their biggest one on the go the moment.
00:10:56
Speaker
First light rail system in Penang, 21 stations, 29 kilometer viaduct, and they're going to build it in six years. How much you reckon that would cost you? viaduct with 21 stations.
00:11:08
Speaker
i tried to do some I tried to do some similar math on some projects, like looking at the volume, if you looked at like relative labor or material costs and you looked at like a company doing 4 billion US equivalent, like Gamuta might be doing and do actual volume of construction work, it probably amplifies to something a lot higher equivalent. don't even know yet.
00:11:30
Speaker
It's, they've got Koska and it's got as well, there's a huge bridge they're going to build over the straight just outside the city. It's under three billion quid, six-year project, which is incredible.
00:11:41
Speaker
First, first metro-like line in Penang as well, which you think it's like a, quite a well-known tourist center. Yeah, yeah. So putting 21 stations in there will be pretty massive for the city.
00:11:55
Speaker
Presumably there's only lines in Kuala Lumpur at the moment. and Yeah, I believe so the Having been there 10, 15 years ago, it was already pretty chock-a-block with scooters. So chock-a-block in translation in non-Australian slang just means like really jammed full of scooters.
00:12:18
Speaker
I think that's a a global phrase, that one. Is it? yeah i never know.
Trunk Tools' AI Platform and Funding
00:12:22
Speaker
Yeah, interesting market, very large, somewhat fragmented, but yeah, interesting takeaway that like some of the big international specialist contractors have also been successful entering and being and sustaining in those markets. Like i think Boig has been there for a very long time along with some of the other contractors.
00:12:40
Speaker
Yeah. um All right, moving on to technology news. So construction AI startup Trunk Tools raised $40 million USD ah in a Series B funding round, bringing total funding to $70 million. all They reportedly showed 5x revenue growth in six months.
00:13:00
Speaker
For those that don't know, they have built a RAG platform. Double clicking on what that means. It's a AI technique that allows you to, you can think of it as AI and search in combination.
00:13:13
Speaker
um So it's called reti Retrieval Augmented Generation. can think of it as like searching over a bunch of things and then understanding it ah in ah in being able to respond. That connects construction documents, drawing specs, ah RFIs, schedules, submittals, enabling natural language queries and questions to be asked of those documents.
00:13:34
Speaker
This is a massive raise and representing an investment that investors are making that this is going to be a... like a billion plus dollar valued company because that that must represent a valuation ah of probably 200 to 400. Something that, yeah, probably 250 to 500 million valuation would be my my ballpark guess.
00:13:57
Speaker
Something that we've spoken about on the podcast before. Yeah, what do you think this means for AI ah in construction, technology in construction generally, Carlos?
00:14:07
Speaker
and And what do you think of the progress from the Trunk Tools team? Yeah, it's um we had a similar conversation a few months back with Document Crunch. So the ability to search, review, extract information at lightning pace and use that information.
00:14:23
Speaker
It's definitely like a space that I think we can all agree is a very useful thing. And we used to burn a lot of hours looking for information. And I think there was a stat I found, which is 30 to 40% of time in construction is just looking for information.
00:14:35
Speaker
Yeah. which is huge. The interesting one with this is it's not built, it's not like a a bit of UI on another model. ah They've actually built the model. So I think the thing I'm intrigued about is how strong it actually is and what it does more than just sort of retrieving information and maybe sort of suggesting prompts on on what to do next or or what to send or what to action.
00:14:56
Speaker
Surely, using other models, people could just spring up intelligent search type applications pretty quickly. um And they've obviously invested a tiny ton of money doing everything from first principles or from scratch.
00:15:07
Speaker
so I don't know that they actually have built an underlying first principles foundation model. A very quick search suggests that they have. I think what they might be referring to is that they built the model of like connecting a bunch of these pieces. to we Yeah, one of us will get fact checked correctly um for sure.
00:15:25
Speaker
Is it a model where they're like, are they learning about as you go, are they learning about your business and project? Is that... yeah Do you know if it's pulling that data in like as it's reading about your project? Is it creating a model of your project?
00:15:40
Speaker
So if we think about the, and for those that don't know much about ch trunk tools, if you think about what the, if you to try and visualize what the product is and the platform is doing,
00:15:52
Speaker
Construction is ah typically um structured in a way where there are all of these different processes or areas. Like you can imagine you've got an RFI.
00:16:05
Speaker
um you' an rfi which is kind of related to a specification because it's asking a question maybe about a specification but it's not directly related and then you have the specification which is kind of related to a drawing because it's referencing something on a drawing or the other way around but it's not like technically structurally connected and similarly that specification might be referenced in a contract but those two things are not connected and so that's generally considered like unstructured data. So there's lots of information, but it's it's not function functionally connected ah to each other.
00:16:41
Speaker
On top of that, that information that is unstructured in all of these different silos actually lives often in different technical systems. So a bunch of it might live in Procore, some of might live in a P6 schedule, some them might live in SharePoint.
00:16:54
Speaker
So you end up with like underlying data that is unstructured and kind of disconnected, living in systems that are totally isolated from each other, which I think drives the underlying problem that you highlighted, Carlos, which is how much time ah construction teams spend looking for information.
00:17:11
Speaker
and something that we've all probably experienced on projects and just the utility of finding stuff, you know, the needle in the haystack on its own is is probably like and immediately appealing.
00:17:23
Speaker
Then if we like think about the platform, the idea is that you connect all those systems that who yeah the information is stored in into this product. It's ah organizing that information and ah ah vectorizing it. so So applying some AI logic to how things are related to each other is probably like an oversimplified explanation.
00:17:46
Speaker
And then that organized unstructured information now can be exposed or lends itself to kind of like mini apps. And so their product has these kind of like mini apps that tie into that underlying platform.
00:18:01
Speaker
You know, they call them Trunk Submittal, which is a little app that does submittal review or Trunk Text, which is like a chat bot that's that's aware of that information. Yeah, and Perform and others, yeah. Yeah, so so it's like this underlying, there's like a couple of these layers. There's the there's the connectors to all of these tools.
00:18:19
Speaker
There's the organizing of all that information that they're they're ingesting. And then there's the exposing of these mini apps. And so, yeah, I think that that model, well, it's very interesting from a technology perspective because ah once you've got all the information organized, you can expose these very...
00:18:36
Speaker
specific use case apps on the end, which is which is a very interesting model and not something that pre-AI companies do
AI in Construction: Opportunities and Challenges
00:18:44
Speaker
a ton of. Usually those mini apps are like totally isolated from each other. If you go to like the Bentley 140 products, not talking to each other problem.
00:18:53
Speaker
Yeah, so the CEA says they trained their models completely from scratch and they didn't use any general purpose AI because they don't believe it works properly based on the complexity and the messiness of construction. So it'll be interesting to see how much more powerful this is than the others.
00:19:08
Speaker
ah The learnings and the the training is specifically for construction rather than like slapping an app on top of a ah general... Yeah, really interesting given like companies like OpenAI and Sam Altman clearly pushing this angle of like the, it's amazing how far a generalized model can go.
00:19:26
Speaker
And obviously there might be some like marketing jujitsu from the... Trump's all seen like the perceived value of the having trained Joe and model. Yeah.
00:19:37
Speaker
so Like internalized IP is quite um compelling. And I'm sure there's an argument to say you're not just surface in every category. You're going deeper. But like having that vertical and really going deeper heavy on one from scratch it'll be it'll be really interesting to compare them and actually see the the difference in performance um for sure sorry sam they'd have to sell like they'd have to go down you think the path of okay you know to this company we're going to create this model that sits around your entire company so each of your projects using it and you're building out that model over the time the
00:20:13
Speaker
Like it seems like the obvious path because, all right, so you've got all my data, but how are you using it to improve over time? Or are we just starting from zero on every project and dumping all these new things in?
00:20:24
Speaker
Yeah, that's that's where I think that the there's like maybe some literary license going on. So yeah, if we think about the pieces to this, there's the there's the infrastructure to connect to a bunch of apps.
00:20:36
Speaker
which is ah critical for them to solve this problem of getting data out of silos. But ah companies like OpenAI i and other model providers are constantly building this ecosystem of things that they're connected to.
00:20:52
Speaker
And then on the flip side, the things that they're getting the information from, like a Procore, If you really boil it down, like even though they list like Autodesk, Procore, P6, SharePoint, M8 as like their integrations that they support, if you picked a single construction contractor, they don't have all those tools that have Procore plus baby P6 plus SharePoint. So a large majority the information probably already exists in say a Procore or an Autodesk.
00:21:20
Speaker
And so you'd imagine both Brocore and Autodesk are building this type of thing and eat right now. So there's ah there's a ah competitive element of speed here that Trump tools. They definitely executed very quickly. They were founded in 2021 and are moving very quickly.
00:21:36
Speaker
you think an organization organization like this could You say obviously Procore and Oracle and others will be doing similar, but there's this space to compete to be the construction AI model that others license like what everyone's doing with OpenAI.
00:21:51
Speaker
Well, I just, I go back to, I think that we might be stretching the definition of like custom built model. i I could be proven totally wrong and get fact checked, but the there's this connector layer. There is the layer of then I'm ingesting all of these documents and then I have to kind of like do some passing of those documents.
00:22:09
Speaker
I suspect in that passing of those documents, there is some off-the-shelf models being used. Where I think that they probably have built their own model, again, this is a total guess based on no actual facts here, is probably closer to what Sam has talked about, and it's how they might expose these ah little like mini apps.
00:22:28
Speaker
So if you think about the the way that like the trunk text Q&A app works versus the submittal one, There probably the need to be able to customize company best practice in how they review submittals, for example, and put that, you know, bake that into the operation of that mini app.
00:22:51
Speaker
I could imagine that the model that they've trained needs to be clever to work out. Am I going and finding this in RFIs, schedules, contracts, change orders? I think that that my guess would be that there's this orchestration model thing that they have ah trained or fine-tuned.
00:23:08
Speaker
And probably there are like general purpose models doing some of the underlying work. It would be a guess. was going say back on the point around the integrations, when you start to play it out in your head, you look at an Autodesk or Procore who have this huge suite of tools.
00:23:24
Speaker
And like you said, they have all the data sitting there. It just feels very easy for them to go and build the same thing and then have a couple of applications that do, you know, these, the submittals, the text, whatnot.
00:23:36
Speaker
Yeah. But then when you start to think about the myriad of customers that don't use or they use one aspect of Autodesk and then they use 10 other tools and they need something to pull it together, it starts to feel, okay, who's doing that?
00:23:47
Speaker
you know And they seem like they're positioning themselves to own that space. There's probably a ton of players that need or would like get value out of that connection from a bunch of other tools. ah So yeah, there's something there's something interesting there, um but they need to work out what's our differentiator ah to what you know the other big players can do or what someone could custom build with a co-pilot, for instance. Martinez just went and built and launched their own you know AI tool that sits over all their company documentations. i don't know if you saw that, Jace called Marty, they've called it.
00:24:19
Speaker
called marty good name it's it's great they've trained it to have this like awkward australian responses like hey mate yep here it is um but you can just go in there and say i'm trying to find fuck off i'm not doing that for you
00:24:36
Speaker
ah it was It was super impressive. They put a little video up of it, but you go in there and say, I need like this you know operating procedure and it'll flick it straight out to you. So you know they're pulling their own company documents, which is the sort of starting idea of this. so And they've done that with Copilot.
00:24:52
Speaker
Yeah. It's like, cause there's the thing that we haven't talked about here is yeah. So you could have like a protocol or an auto desk trying to build exactly this. There's probably just a, how fast can they move versus the trunk tools like protocol? We talked about have a hundred plus people in a data and AI team, like, uh, seemingly moving slower than, um, than the trunk tools team.
00:25:17
Speaker
But on the flip side, yeah, the the problem that a lot of large companies are thinking about when they're thinking about this type of ah answering questions semantically and pulling data from lots of services is how do permissions work?
00:25:34
Speaker
So if I'm a user in Trunk Tools asking a question into this chatbot about, you know, how does this policy work or how do I do this? there is this problem inside of companies is maybe I shouldn't have access to all of the information across the company.
00:25:51
Speaker
Maybe I shouldn't be able to search what's Carlos's employment contract, for example, but other people should be able to search it. Yeah, or even a supply chain accessing works information of others and things like that. Yeah.
00:26:01
Speaker
And so the benefit that like a protocol has is they will, I'm imagining part of why they're slower is that they're building in this like permissioning problem, which is Microsoft have had the same problem. Like they've been moving slower with their equivalent of this because Microsoft are building this on top of SharePoint for sure.
00:26:19
Speaker
And the hard problem to solve is this permissioning, like the question that you ask being bounded to your permission set. um And so I would be very surprised if that is inherent in the data coming across or that you you may have to be configuring this again in trunk tools, which does give advantage back to a protocol, for example, where you're already a user bounded to some scope of this project and this other project or certain roles on there.
00:26:48
Speaker
Does it become really difficult too, where if we think about the actual end users on sites, like using these ah prompts, chats or applications, there becomes this gray area of overlap with the actual tools that are running the job.
00:27:02
Speaker
And well he so yeah like i don't know how unless they like go backwards and actually action things in other applications, you end up in a, ah well, i I was working in this application and it looks like I've achieved some actions or provided some information or or done X, but actually the the actual apps that are running the job are empty and unused.
00:27:26
Speaker
um Yeah, i I think that, I totally, I think that there's these two schools of thought that kind of happened here where where where there's ah there's an argument that says the underlying apps become superseded by the the the chat style interface or the AI interface. Then they have nail UX like big time and then get really, yeah. And then they need a way to action back in those things. So at the moment, yeah in many of these cases, what Trunk Tools is doing is connecting to, say, Procore,
00:27:55
Speaker
and consuming all of the content from Procore, organizing all of that. and then And then like when you ask a question into many of the, into in like in trunk text, they're often not going and getting Procore data on the fly.
00:28:10
Speaker
They've already pre-progressed that data, which i'm I'm unsure on an integration by integration basis, but in many of those scenarios, then it might be very difficult to, you can't take an action back in the underlying tool.
00:28:25
Speaker
Now, i' they probably do this in some at the moment and are working on across all of them. But yeah, until you can action back in the tool, I don't think the underlying system becomes redundant. yeah ah But the control around the action that you're taking in the underlying tool is also a big scary thing for companies. Like an agent can now go and delete a bunch of data or replace a bunch of data. um So there's there's like, they've got very far quickly.
00:28:51
Speaker
There's probably some like chunky problems around permissions and then starting to write stuff that is ahead. i was say, would you get it? Would you buy it on your project? You're on the ground. Would you give it a go?
00:29:03
Speaker
Oh, we talked before this offline, before we started recording about how much money I've spent on every AI tool. So yeah. Yeah. Yeah. A hundred percent. You're just in your office with AI tools. Yeah. One of these on day one immediately just to search and query and and then extract information.
00:29:18
Speaker
um yeah Yeah. I don't know how you can function without that now. Yeah. I'd be interested in people's thoughts on like this versus a document conscious because I think While they both started maybe from slightly different spots, it's pretty clear from the direction travel of the Document Crunch team and the Trunk Tools team, where Trunk Tools have started with just give us all of this information and we'll start with retrieving and finding stuff.
00:29:42
Speaker
And Document Crunch has started maybe on a subset of the information, um and have probably been deeper into the workflow of what you might want to do with that, like reviewing and and stuff.
00:29:53
Speaker
ah What would be your prediction about how that kind of those two peers proceed over time? I think Document Crunch, if you think about, well, just think about UK market and major schemes, I think it's probably the easiest sell because it feels like a slightly more closed market.
00:30:11
Speaker
We're specifically trying to do these five to ten things well, um particularly from a commercial point of view um and formal communication and notices and things like that. So I feel like that's an easier one to get your head around and sort of invest in for a specific use case.
00:30:28
Speaker
Whereas I think this is so broad, I'd be like, like how much are we changing and the culture of like how we run a project and the amount of checks and balances we'll need because it's so broad, none of it sort of tried, tested and and defined. So I think there's a as an element of is it too big to really sort of focus on until others have proven it works?
00:30:53
Speaker
Whereas Document Crunch feels a bit more like you're buying this service for this. We happen to be doing AI in the background with your documentation. The architecture of it definitely feels more like the utopian idea of, okay, like this is where we'd love it to get to, like connect all these data points together and just have this sort of hive mind on and on the project that we can pull things from. But I agree.
00:31:13
Speaker
By the process of ra you know pulling that out and then connecting it to everything for every customer, it feels feels like a big problem to get past and you know start getting down into the permissions who can see what and where.
00:31:27
Speaker
On the flip side though, I'm kind of of the opinion whoever has the most data and has got the means to build the tools and is tapped into the pain is going to win. So you know if they're going to go and connect to everything and they've got the right team who's really looking at, okay, like what can we do?
00:31:45
Speaker
Yeah, 40 million helps. But yeah, go they've got the data. like It's in front of them. if they can If they can get it right, I think um the utopian idea is like is the better idea.
00:31:58
Speaker
um Whether or not it lands, it's sort of up to them if they can build the tools right. and Yeah, think you're right. The user experience good. It's like ah there's an execution part to this, which is like, it does feel like the the the safest bet would be focusing on a tighter set of use cases.
00:32:19
Speaker
Because I think the big unlock here is being able to action things off the back of the insight and go and do that. And you know then you start having an agent that can go and do some of these things for you and say, here's the five things I'm about to do.
00:32:32
Speaker
You tick them and it goes and does. Like you the the deeper you go, I think there's a lot of massive value unlock. And having a tighter set of use cases means you probably get further down that path.
00:32:46
Speaker
um You know, and on a risk adjusted basis. um Whereas this like everything ah play, ah suddenly you need teams, you know, you need your $40 million dollars to build these teams for.
00:32:59
Speaker
I got a but bunch of use cases for ah RFIs, got a bunch of use cases for contract review. I've got a bunch of use and you're trying to like do all of them at the same time. And to a degree, some of it lives or dies on like the quality of that execution. So, you know, if you've got two of the apps, two of these mini apps that suck and two of them are great, what does that mean for your, you know, deployment in the company? Obviously, we know construction fairly well. Do you think whatever Microsoft do will probably be the winner, even if it's not the best?
00:33:27
Speaker
Talking to companies in Oz, they're all they're all getting ready to do what Martinez has done to some degree, Microsoft. They're already talking to them. They're already set up with SharePoints. That's where they're keeping a lot of
Are Planners Obsolete? A Debate
00:33:39
Speaker
their documents. and so Trunk tools have got to get out of this retrieval space and into action very quickly because the retrieval space is going to be commoditized. Because it's the easy space. Everyone's on Microsoft. Everyone uses Teams, even though it's shocking.
00:33:52
Speaker
um i think or if you're Or a big Procore shop. but They'll do it in Procore first. Yeah, yeah, yeah. But an established vendor that they're already in is the easier win i do i think the retrieval thing gets commoditized quite quickly but uh yeah so the the t trunk tool team's like speed is going to be massive because they've got to get from this just answering questions on on information to taking actions really really really quickly yeah we got five minutes by the way all right yeah i could that was super interesting it could have gone uh much deeper but i think yeah great company and great idea and ah real value uh prop so yeah uh wish them luck and and i'd be a customer
00:34:29
Speaker
All right, jumping to topic number three, ah there was a ah post shared by ah ah friend of the podcast, ah Val Matthews, which ah is the host of the um ah Project Chatter, one of the co-hosts of Project Chatter podcast, claiming expensive planners getting paid 200 to four hundred k a year are obsolete.
00:34:51
Speaker
And citing a series of um reasons for that, that vary from 60% of mega projects, skiing plans, schedules, 70% of planning outputs being unused, complicated schedules being created and and ah disputes being the result.
00:35:09
Speaker
a Spicy take, anyone think that the 200k to 400k planner has a shelf life? you can't You can't yell at AI, really, can you? like The plan is sitting there. You sort of pay them. You can go over and yell at them.
00:35:21
Speaker
and it's I didn't think that's where you going to go with that. the Ultimately, the work still needs to be done. so like You're still going to have to make a plan. that You mentioned in the post, he said you know you go into these um lengthy delay analysis, you get these professionals in.
00:35:39
Speaker
but They need a baseline to look at. ah Until the contracts say we don't need a P6, we're not walking away from it. So not even close. Carlos? but Unless there's something's changed since we left construction.
00:35:51
Speaker
The schedule is so important. It's not just the planning team. The project managers all know the schedule. The schedule does guide all high level thinking. I think communication is the issue. It normally sits within a few people and it doesn't go any further.
00:36:03
Speaker
It's hard to imagine a world without the schedules. And with the these schedules are so complicated and difficult. fucking basically hard to use because the apps aren't very good, that you're going to have people commanding ah a decent wage, albeit two to four hundred seems a bit wild, unless that's Aussie dollars.
00:36:18
Speaker
That's Aussie dollars, yeah. vo there For reference, Vals and Oz. So, yeah, complex delivery without these schedules is hard to imagine a world ah without it, honest like based on the current set of technology and and how teams operate.
00:36:32
Speaker
um I don't know how it works commercially without these schedules. like All you're going to end up is, even if you're on your like a cost plus contract, so you're not stressing about change and ah variations from a program point of view,
00:36:46
Speaker
You're just going to have a million arguments over, this is what you should have done and end up in disputes because there was no command. I don't think he's arguing getting rid of the schedule. think he's just saying that the planner becomes obsolete. But he because can you see a world where the schedule is being updated through another tool and who's feeding information into that tool? There's still the science and the thought behind how schedules work.
00:37:10
Speaker
And a lot of people don't get them outside of the planner, the project manager, and a few key individuals. So unless we completely shift how we deliver projects, I can't see that changing in the next five to 10 years.
00:37:20
Speaker
I agree with the underlying points, which are like kind of statements of fact, which is 60% of mega projects exceed plan schedule, 70% of planning outputs not actively used by delivery teams. Pre-six and legacy tools create logic heavy schedules that confuse stakeholders and slow decision making and legal disputes from project delays can cost millions.
00:37:40
Speaker
But yeah as much as I agree with that, and in no disrespect to ah to a friend Val, I think that the ah the headline is clickbait because I don't think there's any logical connection drawn between getting rid of a planner and solving any of those problems.
00:37:57
Speaker
They are very fair points. So just take one as an example, planning outputs. The planner on the job would always know the output that they plan for to drive the duration of an activity. There is no way the engineer knows.
00:38:09
Speaker
like unless they practically go and ask, there's no communication of that to anyone outside of the schedule. As like a way to wrap this idea, I would almost wonder as like a food for thought rhetorical question, whether the era of the to 400K planner is over,
00:38:28
Speaker
is over And that what we might see is the 300 to 600K planner as many of the the number of planners needed decreases. And some of these actions can be done through the likes of what we saw with the trunk tools. And we're talking about taking actions to update back to the schedule.
00:38:48
Speaker
They've got the schedule, trunk schedule. And the real important thing is ah this understanding of how all of it works together, what the overall... You know, this understanding of the map of the schedule, what's important, thinking about mitigation.
00:39:03
Speaker
Yeah, do you just get less planners? Probably that means that there won't be wage inflation because you'll have so many planners without jobs. Like probably doesn't result in increased wages, but a decrease in the number of planners.
00:39:16
Speaker
Actually, I think I've now just agreed with Val. Yeah. Because if you just reduce the amount of planners that you need on a project, then you're going to get ah through supply and demand a reduction in in salaries.
Episode Conclusion and Next Week's Preview
00:39:28
Speaker
Do you see planners getting more involved with other tools? Do you see them coming off and that'll be the start of the transition? They'll start to become experts in these new tools coming in and being able to facilitate visual planning, rolling out better processes on jobs or they're going to be It'll be this sort of old versus new people stuck in the P6 ways not wanting to move. Well, there's kind of the the like dark art space. So like production rates, um that's always just been like ah historical information and data that they're sort of plugging into schedules to try and generate durations.
00:40:02
Speaker
But now that we have more more data, there should be more of this connection and maybe there'd be more focus in those sort of spaces during a live project. So it removes the sort of the sort of estimating side of of scheduling.
00:40:13
Speaker
And interestingly, no one mentioned schedule generation off the back of it. And also Val, if you do listen to it, none of us are salty that you didn't mention Apex in the collaborative platform.
00:40:28
Speaker
We might have been instrumental in that. No, we're definitely not salty. Righto, over time, Sam, thank you very much for joining for the conversation and looking forward to next week, our all in debate across the Apex team. now Thank you very much, everyone, for tuning into today's show.
00:40:45
Speaker
If you did enjoy the episode, please do think about liking the video or following us on your chosen podcast platform. We appreciate your support and we'll see you all next week. Bye-bye.