Why Clashes Exist in BIM At All with Rubens Lage Lopes

Andrea Ciaffi

Hi, Rubens. Welcome to the Digital Construction Podcast.

Rubens Lage Lopes

Ohh, thank you, thank you for having me.

Andrea Ciaffi

In this, in this episode, we sit down with the Rubens large. Lopez, who is a BIM Information Manager at the GDA Agency, and he has spent more than 20 years in construction across Brazil and now Ireland, and now it sits on the client side in one of the largest campus redevelopments in Dublin.

Andrea Ciaffi

Recently, I found one of his... post on LinkedIn. And he asked a pretty simple question that got a lot of people thinking. He asked, why do clashes exist in BIM at all? And is the answer better clash detection or better rule enforcement? So today we are going to talk about one of those topics that we often discuss here at the Digital Construction Podcast, which is coordination.

Andrea Ciaffi

We know it's about people, it's about software. We'll also talk about clash prevention instead of clash detection, or maybe also clash detection. But it's interesting to think about it from the prevention point of view. Also, we, Rubens, will talk about information management because the project is working on right now.

Andrea Ciaffi

as an interesting aspect. And we'll also talk about AI and where AI fits in all of this story. Rubens, before we start, I have an icebreaker for you. I want to ask you, what is the best piece of advice you have ever received in your career?

Rubens Lage Lopes

Okay. Ohh. Yeah, I think. It was to kind of go slower in my case because sometimes I want to do things too fast and I have like a lot of ideas and so to have more patience and go like try to pace myself.

Andrea Ciaffi

Yeah, so, yeah, it's interesting because we live in a very fast environment in especially in construction with very quick deadline, but yeah, the idea of slowing down sometimes, it's a good approach.

Rubens Lage Lopes

Yeah, I try.

Andrea Ciaffi

And... It's a journey, it's a journey. So Rubens, you spent more than 20 years in

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

construction and you built a career across Brazil and Ireland. When you moved between these two different industries, what did you learn about these two different industries in two different places? About Beam.

Rubens Lage Lopes

Yeah, like in Brazil, I start start studying civil engineering technician in 1988. And then I started working, was always on site. Like I went from school, college, to like a work on site. And then I saw those, what now I know, clashes, but they're back there, conflicts and problems.

Rubens Lage Lopes

It was something like. people think was common. And was everything was like too rushed, the design phase. That thing that I don't think is much different than here. The only difference I see, like a big difference when I see when I move to Ireland, it's like here.

Rubens Lage Lopes

people are trying to, like, we are trying to do better. Of course, they're trying to do better as well, but they have more resource, they could have more bigger projects, and then I start working in bigger projects, and then I saw, like, okay, there are more, say, more money available to actually.

Rubens Lage Lopes

Trying to do the right thing.

Andrea Ciaffi

Interesting. And I know you are interested in information management. And when you look back in your career, was there a specific project or a specific moment where you thought, okay, it's more about information management is not just geometry? in Beam.

Rubens Lage Lopes

Yeah, like, when I was doing, when I was doing my masters in being, I ran one of the professors, he's saying like, oh, the contract route, like the procurement route is very important. And I was like, like, what's that?

Rubens Lage Lopes

Like, That's ********, no, and then, and then later on, like when I start working with like data centers and very fast-paced projects, and then I start like really feeling the difference between designing 10 and designing build and all those different procurement routes and how that was affecting.

Rubens Lage Lopes

the way people were thinking and doing things, that's when I realized actually, you know, the steps we have before, the information and procedures we have before affect the way we're going to work and the outcomes. So like, that's why I started thinking, yeah, like we need to, the information.

Rubens Lage Lopes

it's the coordination is a consequence of the information we have before.

Rubens Lage Lopes

Also, like I had one post, I remember once I was suggesting we have like a design coordination rate, something like, you know, the bear rate, you go to the building to like, or produce, and we have A, B, C, D, F. And then I would suggest, oh, maybe we could have a design coordination rate, so at least we could answer a simple question, say, like, is this design ready for the next phase?

Rubens Lage Lopes

So I posted that, and then I received a lot of like comments and a lot of like comments against that as well, like, oh, it's too much simplification and stuff. But one of the very, very important comments was Raf from the Arc Docs.

Rubens Lage Lopes

He said, like, oh, you can't talk about coordination if you don't have the information there, the level of information need there. And then that was another thing I started to investigate. So those two things, like. The procurement route, like when I was in college, that was a big moment.

Rubens Lage Lopes

And then this post and Raf pointed that out. I was like, yeah, actually, we need to take that in consideration.

Andrea Ciaffi

Yeah, pretty good approach. And I know you're passionate about coordination. And we know coordination is about, it's about people, it's about software. But when you walk into a project, how can you tell whether the coordination process is working or not?

Rubens Lage Lopes

Yeah, I, the first thing is, like, if we have a process, you know, it's not just people jumping into the meeting and then throwing models on top of the other, like, I, I followed this follow process, like, first we need to define what...

Rubens Lage Lopes

good coordination means, like what well coordinated means. And then to get to that, we need to understand, okay, what kind of project is that? What are the priorities? What design phase are we? So if you're in the concept design, so of course you're not running clash detection all the time, you're not like a. ..

Rubens Lage Lopes

You, you need to 1st care about the concepts with concept design, and then you know, so if I see the project, we have a process and everybody's clear what are the goals and targets and what the standard, what's acceptable and what's not acceptable.

Rubens Lage Lopes

at each design phase. Yes, I think, okay, we are heading to a good process. We had to do good outcome, but otherwise, like, it's going to be a mess and back and forth.

Andrea Ciaffi

Yeah. And sometimes, you know, like in projects, coordination is often treated as a software problem. People fighting between which software should we use to coordinate this specific project. But probably feels like it's more like a people problem, right? People and maybe accountability problem.

Andrea Ciaffi

Why do you think that people still confuse these two, like the software versus people process?

Rubens Lage Lopes

I, I, I think I, I used to do that as well, like, because I like the technology, so I want to solve everything with technology, so you know, I had a little piece of software, there we go, that's gonna solve my problem, then, then Camera AI, that's gonna solve my problem, but...

Rubens Lage Lopes

Actually, like we were talking before here, like we need the three things. We need people, procedures and technology to support that. So I think I remember that time for me it was easier to talk with the software than talk with people. So that's why I was trying to solve everything using the softwares and then trying to find the best software, whatever.

Rubens Lage Lopes

But actually, now I think it's easier to talk with people, then we get alignment, and then the software is just a consequence. Sometimes even crap software can solve our problem if we have good alignment.

Andrea Ciaffi

Yeah, and I want to talk about your LinkedIn post, because in your LinkedIn post, you put it very simply. You said,

Rubens Lage Lopes

Mhm.

Andrea Ciaffi

in the real world, two objects cannot occupy the same space. Our models allow it because every discipline's design in its own model.

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

Why do you think the industry has accepted that as a normal? For a long time.

Rubens Lage Lopes

Yeah, like I had this idea, like I was, I was in the busy building, we saw a lot of pipes and ducts, you know, like it was an office, no ceiling, you could see everything was supposed, and then I was looking, it was like, you know, there was zero clashes here.

Rubens Lage Lopes

And then, and my computer had like a model full of clashes, and then what's the difference? Like, because in reality we have laws, the physic laws that do not allow two things in quite the same place, same space, so that's what came the question, say, why do we have clashes in beam?

Rubens Lage Lopes

And one of the persons in that... post, he answered that. He said like he works in Bing, I don't know. Since, I don't know, 80s or 90s, something like that, and then he said that the beginning was a software constraint, like people it was too heavy for a software to actually block all those things, and you know, to block one thing to clash to another, so...

Rubens Lage Lopes

They just software has to be developed like that. So there was, okay, you do whatever you want and then we figure out later because that's what we have. But since then, I think we are not questioning that. Like,

Andrea Ciaffi

Mm.

Rubens Lage Lopes

why are we still allowing the new one model things, like absorb things, like... with no rules. I think it's just we can think more about that and be more critical and then find a way. So what I'm what I'm proposed, I'm suggesting is like first I was calling coordination rules.

Rubens Lage Lopes

So for example, and I have, I use that in my projects. So at the beginning, we decide, okay, is a precast structure, for example. So then the structure engineer say like, we cannot have any opening in our beans. So that's. That's a rule, so you will not pass any pipes there, because if you pass, you're gonna have to move later, so it's real work.

Rubens Lage Lopes

So, I call this coordination rule, but then people say, like, ohh, rule is it sounds like a and no exception, so now I'm changing a bit to say, like, okay, coordination. principles, okay, but landed there. So we define a couple of them. Let's say we define like 5, 6, 10 of those.

Rubens Lage Lopes

So then when you're modeling, if we know you cannot pass through the beam, so why your software let you do that? So why not we bring those rules from the conversation into the software?

Rubens Lage Lopes

We have technology for that already, but I think just people are not like trying to connect the dots yet, so I, I'm trying to connect those dots now with an add-in, so like you create a rule and then the add-in Revit will block or create a warning, say, like, are you sure?

Rubens Lage Lopes

You, if you're breaking that rule, we agree there, so...

Andrea Ciaffi

And you'll do that directly in Revit, for example.

Rubens Lage Lopes

Yeah, so what I'm doing is it's a web application, then the BM manager can use this web application during the coordination meeting, and then they discuss, he goes creating rules or principles, and then say, like, and... They create the rule and create the consequence.

Rubens Lage Lopes

So the consequence is block, like literally when the guy is trying to create the rule, the software are going to block the action or a warning. It's going to pop up, say like, look, that's coordination rule. Thirteen, you cannot pass through a concrete beam.

Rubens Lage Lopes

Are you sure you want to proceed? Yes, I want to proceed. I don't care, but then you pass through, but at least it was clear, like it was not unintentional, you know.

Andrea Ciaffi

And so you mentioned, is it the BIM manager that sets the rules or principles?

Rubens Lage Lopes

Yeah, I think, yeah, the B manager would set that based on the B meeting. So you are in a B meeting, everybody knows, and we get to an agreement, and then, okay, set the rules. And, but even if the B manager set the rules to block.

Rubens Lage Lopes

Whoever is modeling can log an exception. So I need an exception here because we

Andrea Ciaffi

Mm.

Rubens Lage Lopes

need to discuss, for example, what if that beam, it's blocking a pipe? Actually, the pipe has nowhere to go because there is a ceiling just underneath the beam and the beam is still too high. And then say, I can't, I have to pass to here because I don't have other option.

Rubens Lage Lopes

Or if like the pipe was there and the beam came later. You know, that's another problem as well. So all those nuances, different situations, they need to take into consideration. But I think not because it's complex, just because it's complex doesn't mean that we should not deal with that.

Rubens Lage Lopes

I think we should. Like we must deal with that because we can avoid, if we avoid 30%, 40% of rework of this like silly mistakes or things that we agreed and then people are just not following, I think it's a huge game.

Andrea Ciaffi

Yeah, and when you work on, so I imagine there will be, sometimes you have a B manager per discipline, right? Because different companies work together, different disciplines work together.

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

In that environment, do you see like discretion of rule and principles more, there is one B manager that creates the rule or is it like Is it like a coordinated process between the B managers of the different disciplines?

Rubens Lage Lopes

Yeah, it it's funny, like I'm calling being manager, but in other company could be a being coordinator,

Andrea Ciaffi

Mm.

Rubens Lage Lopes

and what you describe that one being manager per discipline from my in my terminology I would call being coordinator or sometimes discipline leads, so.

Rubens Lage Lopes

Companies that have many different names, but should have one that's leading the project, even if you have one per discipline, one per company, and at some point we would have one that is like a, like a, it's the one it's really ruling them all.

Rubens Lage Lopes

So I think this one should coordinate with the other coordinators or the other leads to say like, okay guys, what we agree here. And then they get to some agreement. And then these agreements become a BEP appendix. So it's a document and papers that people can look at.

Rubens Lage Lopes

And, and that's the difference. I, I say, like, for example, I was calling like a reactive coordination, preventive coordination, and proactive coordination. So, the reactive, it's like what people everybody does, like, we model and then we have the problem, the clash there.

Rubens Lage Lopes

The preventive will be this documentation, where we agreed rules, we agree principles, and then say, like, guy, when you're modeling, remember, look at this document here, and then follow this. That's what we agreed. And but I think the proactive is this, like we have.

Rubens Lage Lopes

An application that turned those rules into... actually actions in the model. So yeah, I think we need to have one that's on top of the others and then set this.

Andrea Ciaffi

Yeah, it's good to have a framework. And changing gears, talking about information management. I know that you're working on a project

Rubens Lage Lopes

Mhm.

Andrea Ciaffi

that it's a long-term public redevelopment and the information that lives inside this project has to outlive the project itself, right? It will stay there for quite some time. So, when you think about information management for a project that this is like a it's a campus, I think, how do you how do you think about information management when there is such

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

a long horizon for the data that needs to live there? How do you change the information requirements from day one?

Rubens Lage Lopes

Yeah, it, it's it's the company is kind of a meeting city that it's running like, I think it's 20 twenty-five years since they they create the agents, and of course, like... Rules are changing since then, systems, and then so we have legacy information, we have new information, new projects, and the new projects need to get the old information and bring it in.

Rubens Lage Lopes

So the information management there, it's very complex. Because, like, we need to deal with old database, new database, new standards, even naming, you know, so do you have old naming, new naming, new naming convention? It's... And it's hard, one of the top problems there is like, it's not that we don't have the information, it's like how to find the information.

Rubens Lage Lopes

Because this is different places and different formats. It's, yeah, it's challenging. Sometimes it's. It's boring, so much documentation and so much procedures, and then sometimes, like, I need to fight for like silly things, like, ohh, naming, ohh, you name this, it's a bit different than this one, and then to know, like, people are so busy with other things.

Rubens Lage Lopes

But then you, you're fighting for naming, you sound so like, ohh, jeez, like this guy again, but like, it's yeah, it's a very important, and because otherwise, like, we can't find things, then we want to kind of duplicate information or take wrong decisions.

Andrea Ciaffi

And. And yeah, another interesting aspects about them in convention, how you organize a project is ISO 19650, which is another beast that many people have to deal with.

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

Do you find do you find that this standard is applied in on live projects like we expected? On paper.

Rubens Lage Lopes

No, no, I don't think it's the whole thing is applied. And, and then that's my 2 bits here. Sometimes I think it's too complex, but so I think one of my my one part of my job is to translate all that complexity and complication to the into something simple that people can just act on and say, "OK, I understand why you're doing this."

Rubens Lage Lopes

Ohh, why we have those folders? Why do we have suitability code? Why we have versions? Do you know why we cannot I should not put the date on the file name or then create folders with dates? Why do I have to overwrite and then the same file and then create versions on top of the other one?

Rubens Lage Lopes

So those kind of things, I think it's part of my job. It's to translate and bring that. But I don't think the project it's using, I think each company, they have their own pace, their own priorities. So they're, I think, being managers across.

Rubens Lage Lopes

Projects that they're trying to do their best. and bring the best as possible. In the public jobs, it's a bit different. because we have a legislation, so there are other standards that have to follow. So I think we follow more tightly, and then maybe that's where people also think it's more bureaucratic.

Rubens Lage Lopes

But yeah, we try.

Andrea Ciaffi

Yeah. And, and we can certainly we can certainly have an episode about beam and forgetting to talk about AI, right? Because we are in 2026, it's the it's the elephant in the room.

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

There is a lot of noise about AI and construction now,

Rubens Lage Lopes

Mhm. Yeah.

Andrea Ciaffi

but from your point of view, from your side, have you see... Have you seen any decent application of AI inside Beam that is not just like fluff or?

Rubens Lage Lopes

Yeah. Yeah, no, I, I use it every day. Like, I started using, I start programming, programming and Dynamo and Revit, and then I remember when the ChatGPT

Andrea Ciaffi

Mhm.

Rubens Lage Lopes

start like launching, and then... I will say, okay, maybe I could try to get the chat 50 to create a Python script to run into my dynamic script and then start working. So, okay, so, and then from that I start using AI to create web applications, a simple one.

Rubens Lage Lopes

For example, we had in our my previous job, we need to check a lot of file names. And then see if they comply with the naming convention. So then I got, like, I got an Excel and then the ChatGPT and said, like, okay, get this Excel turning to, at first was like a scripting Excel and then become a web application.

Rubens Lage Lopes

So answering your question, I think. It's easier to do to deal with huge like big amount of information because of AI, so that's what the way I'm using now. So, the one that like a typical design package, we would have 400, 800 files like.

Rubens Lage Lopes

that had a practice that was 1200 files. I could never check all the naming convention, check if all the files match the deliverables list, and then checking if the models have all the metadata that we agree. So all those things, it's, they're very like a mechanical, let's say mechanical tasks.

Rubens Lage Lopes

So, that I did faster because I used AI to help me with the code, and then the code create this all automation to do that. I think if people are using like that, so it's not slow, but if you're if I using... AI to do my job instead of do the thinking for me and that's a problem then.

Andrea Ciaffi

Yeah. Definitely. But speaking about thinking, I think about your idea about rule or principle enforcement. It feels like one of the things you could

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

potentially do with AI. Do you feel like maybe the creation of these rules or maybe potentially helping predict where Clashes could happen before they even happen. Do you see any any role for AI apply to the to your idea?

Rubens Lage Lopes

Yeah, that's it, so I developed the the adding that goes in in Revit, and then then this web application, so the the beam manager create the rules, but the the next step is that I'm developing now is like a get AI to read.

Rubens Lage Lopes

To BP. and then suggest the rules. So help the B manager. Or if you have a B meeting, then you can bring your B minute, and then the AI say, based on your B meeting, what you agreed there, the rules would be those.

Rubens Lage Lopes

And then the B manager say, yeah, this makes sense, this is not, they let me change here, here, and that. In terms of predicting, predicting clashes, I think if we have the rules, you have the principles, then the AI can start like working on those things and then start predicting.

Rubens Lage Lopes

And I think we can go beyond that is like instead of we talk about. clashes. Let's talk about money. So everybody cares about money. So if I say, if I say to my, to my, that's what something happened, like what the director came to me, said like, I don't care about clash movement.

Rubens Lage Lopes

I want to know if the project is ready, what is priority? And then I said, okay, so then we set. rules to define what's high, medium, and low priority. So then AI can get, if I have the rules, AI can go in the model and say, look, you have 200 high priority, 3,000 medium, 5,000 low.

Rubens Lage Lopes

So then we skip all those noisy, we go straight to the 200, that's high priority. And then that's a, we're talking about money, let's say, then if I get AI to say, okay, I give you the project location, the project type, and those like cost databases, last estimate, what if those clashes, those 200 priority clash goes to construction site, how much?

Rubens Lage Lopes

Cost exposure are we talking about here? So, and then if the project, let's say, projects like 200 million, for example, and then we have, like, I don't know... 100K risk exposure, so that's fine.

Rubens Lage Lopes

But if it's the the if you if you're talking about 20% cross exposure, so that should be enough to go to my director, go to the client and say, like, we need more time for coordination.

Andrea Ciaffi

Mm.

Rubens Lage Lopes

Because if you passed on to tender, you're exposing yourself like to 20 million risk. You know, so, and then that's when they hear when we talk about money. If we talk Beam, no one here, no one care about Beam. Yeah,

Andrea Ciaffi

The listener.

Rubens Lage Lopes

no one, nobody cares about me, they care about money. So let's talk about money. Let's translate.

Andrea Ciaffi

Yeah.

Rubens Lage Lopes

Being to money, yeah.

Andrea Ciaffi

Yeah, currency. Yeah, makes sense. And thinking about the future,

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

it's always interesting to me. Of course, nobody can really predict what's going to

Rubens Lage Lopes

Mhm.

Andrea Ciaffi

happen. But when you think about the next decade

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

and what shifts in the way we design, what's going to change in the way we design and coordinate buildings? Is there anything you think that people are underestimating at the moment about where this the future will what the future will look like?

Rubens Lage Lopes

Yeah, I think we still working a lot in the old way, you know, we use PDF, but it's not a PDF, it's just a digital version of a paper drawing. Like, hey, why don't we build directly from the model? Why do we still using drawings?

Rubens Lage Lopes

There are cases like the hospital and

Andrea Ciaffi

Mm.

Rubens Lage Lopes

Norway that they get no drawings, a bridge, and I think it's Sweden they get with no drawings. So I think the future is like more and more being professionals, they become more proficient on. Using AI to automate tasks to read the data to be more like then we really get the results being and having like better quality because and go beyond the clash detection.

Rubens Lage Lopes

And then actually analyzing level of information need, level of detail, and then models actually become meaningful. Like, I this week I got a quantity survey and complained that I can't take, I can't use the model for quantitative takeoffs, for example, and he was not complaining about my project, he was complaining.

Rubens Lage Lopes

The industry in general, and why is that? It's because models are really, really ****, like they are like the geometry looks nice, but then the metadata, the level of information, the detail, it's crap, and I don't, I'm not gonna blame the... The beam managers, the beam coordinators, they're very busy already modeling and trying to coordinate the geometry, so that's the future.

Rubens Lage Lopes

I think like we need to bring those tools to help us with this extra layer of like data data analysis, because I see the beam manager as a big. Coordinator is a is a. data analyst as well. But to do that, they need to bring AI and bring programming.

Andrea Ciaffi

Yeah, yeah, because these people are also struggling with the deadline and trying to be on time. Yeah,

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

sometimes the big models are not ideal, not because people are not capable,

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

but sometimes because people simply don't have the time or enough time allocated to

Rubens Lage Lopes

Yeah?

Andrea Ciaffi

it. Yeah.

Rubens Lage Lopes

Yeah, and and another problem is they, they, if they had they find the time, let's say they work extra time, we can and they solve it, it's not valued.

Andrea Ciaffi

Yeah, through.

Rubens Lage Lopes

So that's where I think we need to talk about money. So you need to do one more step. Like if you solve a big problem, don't go to your manager saying that you reduce like 20,000 clashes to 200 clashes. They don't care. Say that you reduce the cost exposure from 20 million.

Rubens Lage Lopes

To 200K or 150K, that's what they care.

Andrea Ciaffi

Yeah, that's a very valid point. And I want to take the opportunity now that I'm here with you, very senior person in this job and with all of your experience and try to think about all of the young people that are entering the market at the moment or they're preparing themselves to enter the market in the moment at the moment.

Andrea Ciaffi

You know, it's a quite quite challenging time because it's not just about, I remember like many, many years ago when I was learning about this, you learn the skills, you improve doing the jobs. But now maybe it's a little bit more scary because there is AI and people wonder, will this job still be there?

Rubens Lage Lopes

Yeah.

Andrea Ciaffi

What type of, I guess the question is, what type of skills or mindset do you think this Young being professional will need to have to enter the market and then stay relevant.

Rubens Lage Lopes

I think, I think if they... They understand the process and that we need people to do the process. So then we define the requirements, we get to agreement, we get people on board, and then in the background, they are using the technology to solve this.

Rubens Lage Lopes

Data problems, and then turn those decisions and actually something practical and actionable. I think that's a big skill, so learn how to talk with people, how to lead, and then bring people together. In the background, learning how to use programming, how to use AI to get a BIM model that somehow it's a database, and then analyze this database and give like a real business insight.

Rubens Lage Lopes

Forget BIM, don't go too much into. Tobing. Need to talk. Talk business, talk money.

Andrea Ciaffi

Very good advice. Rubens, we are at the end of this episode. If people want to reach out to you, where can they find you?

Rubens Lage Lopes

You can find me on my LinkedIn. I have two websites, coresite.com, I have valueblocks.com, where those solutions that I was talking that I developed, now they are available if the person wants to use, go there on the website. So, yeah.

Andrea Ciaffi

That's nice. Rubens, thank you for joining us on the Digital Construction Podcast.

Rubens Lage Lopes

No, thank you. Thank you for having me. See you.