DESIGN WORKING GROUP MEETING 30-09-2025

GEORGE STEVENSON …was held accountable on a particular project because they’d taken on the responsibility for as-builts, or making sure that as-builts were delivered, when they clearly didn’t have the scope or the authority to do that. The long and short of it is that if you go through the spreadsheet as this guy was doing with me, he’s quite right that they’ve taken out, there is no Regulation 38 in there. So we think that that probably needs to be addressed. Is anybody else doing principal designer building regs role?

ANDREW DE SILVA Joe, do you have any experience from your team or your colleagues?

JOE STOTT Not that I’m aware of. I think we do have a couple of people in the practice that are doing the role, but how they’re doing it I’m really not sure.

ANDREW DE SILVA I think it will be useful, George, if you could maybe link me to that person that you spoke to because then I can talk in more detail with them. My initial gut feeling is that it may be that things like Regulation 38 and some of the others, because that tracker was predominantly to do with the relevant regulations. So I know when Dita set it up he was looking at it from that point of view and there was a reason why not every single one is in there. Whether it’s to do with what you noted about an RIBA member or something or whether Dita had a different filter in mind when he was setting it up is something..,

I’ve emailed Dita a couple of times about it so it will be useful to have that smaller discussion with the gentleman that you spoke to. And we can also ask Dita directly because to be fair, he did it under his own steam, I think under more the Eric Parry Architects. And then I can also ask there is a gentleman from Stroma who is helping us, advising Waites on the project that we’re doing. Because he’s been a previous building control role for many years and he’s also advised on the Building Safety Regulator. And he really does like his legislation and understands the legislation really well.

So I can ask both parties because it will be really important if there are things that are missing. So if you wouldn’t mind emailing and copying me in to that other person that would be great.

GEORGE STEVENSON What I’d prefer to do because he’s a consultant himself, so I’d like to introduce you to him and you can have a conversation. But he’s a bit frustrated that RIBA have just provided a spreadsheet and now everybody has a tick in the box that everybody can follow without actually properly understanding what the detail is. Now whether he’s right or wrong I don’t know, but he’s very knowledgeable.

ANDREW DE SILVA Franco, thank you for joining. Do you want to let us know a little bit about yourself?

FRANCO PITTONI I’m speaking from Florence in Italy. I’ve known George for some time, we’ve had many discussions on data and this kind of stuff. This is the first time I’ve participated in this group, so for the time being I will be mainly listening. I am a mechanical engineer and I have worked for many years in project management, project controls. And in the last decade I have become particularly interested in data management. I wouldn’t classify myself as a data scientist, but I have suffered from the lack of structure in managing data.

One of the words I have pronounced most frequently in my life is common language, an obstacle I have encountered in many situations in many countries in many projects. I worked almost all of my life overseas, so in my almost 50 years of experience I spent 45 years overseas in international projects. So I’ve encountered the kind of problem I mentioned in many countries, so I think it is a real problem to manage data properly.

ANDREW DE SILVA And James Dixon, it would be good to know a little bit about yourself as well please.

JAMES DIXON I have been working in main contracting typically for the last 13 years. Prior to that design side of clients and consultancy, but I suppose the vast majority of my career has been spent main contracting now coming through the design management function. I’ve spent a long time looking at process that goes within that discipline. Design management in the UK is very significant and an important role given the design & build nature of contracts. And to disseminate what’s good and to place that in the core view of site teams.

My role is developed and my interest is developed beyond what good design looks like to incorporate. Actually, where are we with design in terms of progress? I see a lot of opportunity for improvement in terms of productivity when it comes do design. In construction my experience, we plan very carefully how much concrete, how many doors, how many walls we will erect, all of that good stuff. I think that the planning of actual release of information could and should be better and data helps us to actually track where we are with information release.

In the last 3-5 years I’ve spent a lot of time trying to help project teams actually keep them on track. I'm often on call for when's the schedule being released? When's the door layouts being released and wall layouts and the responses, and hopefully by the end of the month that sort of response wouldn't cut much mustard if we were in a construction world when we would ask the HVACs, are they being installed on Friday? It's much more binary, whereas in design we seem to be more accepting of a loose commitment whereas I’m trying very hard to actually…

And there’s a lot of it that’s on us, we need to know when our procurement is what’s long lead and therefore what information do we need just to manage it. So I suppose that’s a little bit deep dive as to why I’m interested in data.

ANDREW DE SILVA Pierre, do you want to give a quick overview? And then Joe and then I’ll introduce myself.

PIERRE SAUNAL I’m an architect by training, but we are a company of digital and building information modelling, so we create the digital twins in many countries. Our experience span the UK originally, but we are involved now in the Middle East, mainly Saudi Arabia and North Africa, with Morocco, Nigeria and Senegal and have worked on every continent in different degrees for all the architects, engineers, construction company. So we have that experience in digitalising the built environment.

And we are pushing now to the next step because we started earlier this year a new start up to automate the Scan to BIM process from point clouds, creating an IFC model automatically from the point cloud data generated by a Lidar scan. So in a nutshell that’s what we do. George, thank you for bringing us onboard.

GEORGE STEVENSON One of the things that’s just worth mentioning that you’ve not said there Pierre, is that we were brought together by SOCOTEC, they are major building control people. And they’ve brought me, as Active Plan, together with Pierre and we’ve been exploring something that’s joined up there.

PIERRE SAUNAL Yes, it also comes from an experience we had with one of our buildings having to go through Gateway 3 with the new Building Safety Regulator. We were one of the first two schemes in the UK who got granted Gateway 3 approval with a scheme which wasn’t supposed to go through Gateway 3, completion was delayed onsite. And one of the reasons why we managed to go through was because of the BIM model we developed.

So we had a BIM model as-built before even we went onsite, so we were able to demonstrate that during the design process we put in place all of the measures which were required. We were able to show the differences with what has been built onsite, update our model and use the as-built to show that even if we were not structured to go through Gateway 3, the process we put in place from a design, BIM and construction perspective complied. So that has been quite a journey, but that through us way closer to the issues around Building Control, principal designer, be it like CDM or building regs.

ANDREW DE SILVA That’s really interesting, Pierre. Was that a project in Wembley?

PIERRE SAUNAL No, it was next to Limehouse in East London, a 52-unit scheme which had the complexity as well to have been abandoned by a previous contractors. We had to deal with existing structure on site, existing work partly done, a listed facade that the Council wanted to retain, which we modelled from point clouds. So yeah, we had a bit of everything on that project and hopefully arrived on top of things.

ANDREW DE SILVA It sounds like you’ve gone through the difficult route which actually means you’ve probably learnt a lot at the same time.

PIERRE VENTER Architect by profession, not registered in the UK. I’m a BIM consultant interested in data, interested in the data for management over the lifetime of a building, from the AEC plus OO, owner operator. I’ve been in the Middle East, I’ve been in UAE, Saudi Arabia. What’s really important for me is I’m doing a presentation on Thursday at the BIM Coordinator Summit in Ireland on performance matters. And it’s basically two parts. The first one is performance is important, but it’s also the things that make up performance.

And I think all too often in the discussion that I’ve heard so far is that we don’t pay enough attention to actually measuring a benchmark, having a look at the process and then what we get out of it. So James, in a way it’s tying in with what you’re looking at, is that the data is available, but who owns it? Who nurture it, who looks after it? And George knows well that’s what you’re going to do at the end of the day, it’s going to become the asset management. So my interest is in all of those things, architecture, design and the technology etc.

JOE STOTT I’m an architect by training, but my current role is a BIM manager for AHR. We’re a company of architects and building surveyors. We have quite a strong Scan to BIM team in-house, so I was quite interested to hear what Pierre was saying there about trying to automate those processes. But over my role as a BIM manager I’ve seen the transition from 2D AutoCAD, I’ve seen other transitions prior to that as well, but from 2D AutoCAD into 3D Revit and BIM modelling and that’s been great. And we kind of tick that box in the design team we’ve all got these great efficiencies, but I’ve always been interested in what next and how do we shift the focus away from geometry, kind of Revit and BIM being about geometry and automating drawings. That’s all great, but how do we shift the focus towards the data?

Because I’ve always had this personal challenge of how do we weigh or measure the quality of BIM data? I can have two models that appear to be identical, one of which is useable and the other maybe not so. So I’m really interested in how do we shift that, particularly from an architect’s perspective. Because as an architect I think I can say this safely that creative architects are great at design, but maybe not necessarily data focused. That’s not necessarily our strong point as architects in the industry. So yeah, I’m learning all sorts from this group and all of the other workshop groups that I’m involved in.

ANDREW DE SILVA I’m Das from David Miller Architects. We’re a practice of about 20 in Liverpool and London. I’ve sort of taken the role of technical director, but also interested in the CDM and the Building Safety Act principal designer roles which we offer across different projects, across different sectors as well, I’m working with and have known George for 12 or 13 years and it sort has grown into talking about BIM in housing, BIM4Housing, and trying to find the practical solutions that designers within the design working group can do, need to do, can follow, to navigate various regulation requirements.

But also as the team here, there is a lot of BIM and data discussions as well. And I think they’re coming together because historically in the practice when I first started 15 years ago we moved into using just Revit as a tool and then wanted to always stay ahead, look at what the next thing might be. And the reason for using tools and digital workflows has changed from one of efficiency, one of collaboration, and now I think to one of statutory requirements, post the kind of contractual requirements that came in 2015. It’s interesting how that kind of almost the stick and the carrot has changed a bit.

We are now having to use digital tools, the data outputting them to simply make sure that what we’re doing is compliant. We can demonstrate it, we can evidence it and we can audit it. It’s really quite interesting to see the team today because it’s international, but all of you had input into some of the Gateway stuff as well. And what I did in the last meeting was just give an update on a project which on the face of it is quite straightforward and it’s around the Gateway process. It’s a scheme where we are re-cladding a number of blocks in a site just by Canary Wharf. But the building was built something like 25 years ago, so not very old.

But it’s interesting where the existing, Waites are the contractors on that scheme and under what’s called a developer’s pledge they’ve gone back and are looking at schemes that they’ve done in the past which may need remediation work. Not necessarily nothing to do with the fact that they wouldn’t have been compliant when they were built, but more because there is obviously things like timber cladding on this particular scheme which they are going back to remediate under that developer’s pledge. So, it’s a first project that we and I’m going through, through the Gateway process. And timeline wise we’ve gone through pre-app process.

We’re about to go in for planning in a couple of weeks’ time to remove some of the timber cladding and the timber decking to balconies and replace that. So it’s not contentious as such, which is quite a good thing for a pilot project. But to be fair to Waites they’ve appointed various specialists to help and advise the team, including facade consultants, a gentleman from Stroma and other registered building control approver to advise the team. A structural engineer, a fire consultant and ourselves. So we’ve got a nice team. We go for planning in a couple of weeks with Gateway 2 submission at the end of January and then we see how long that goes.

I think a lot of you involved may already know there is quite a long delay from the statutory period or the target period that the Building Safety Regulator has put to actually getting approvals through the Gateway. And I think this group and other groups are trying to share best practice so that we unblock that kind of issues that are happening because it will benefit all of us. So the more of us that learn from each other about the information required, how to submit information, how to make it easier to understand, the quicker we get the backlog clarified and the more projects can get to site which I think is what we all want to happen. From the last meeting, this feels quite analogue in a way, but it’s just sharing some of the things, the trickiness that we’re going through as well.

GEORGE STEVENSON Das, could I just say something about what you’ve just said there? I just wanted to reinforce the importance of the data that Joe was talking about earlier and how it all hangs together. The reason that I’m mentioning this is that as Active Plan we’re doing audits of schemes that have been delivered. And what we’re doing is we’re looking at things that are likely to get picked up, like the fire & emergency file. And some of the basics that we thought would be reasonable to expect, like door schedules having the same references as the drawing that’s being provided, which I think is a fairly reasonable thing to expect, it’s quite shocking that we’re finding that isn’t the case.

What we’re also finding is that even the COBie data isn’t consistent with what has been in the model and it’s certainly not consistent with the schedules. Now, I’m just mentioning that because the whole point, it’s a complex enough thing for building information to be provided anyway. But if we end up with them all fragmented and people have gone in and manually edited them, for example, it’s chaos. So, one of the things, the whole point really of what we did 5 years ago when we set up this BIM4Housing group is really to look at what information is being passed from party to party. And try and look to see how can we make it so that, for example, you as designers receive from developers is structured in such a way that you can use it easily.

And correspondingly, if you’re passing information on to contractors or passing them on to asset teams, it’s important that that comes through as well. And one way of doing that is making sure that it’s all machine-readable. On a project that we’re working on at the moment, we needed to check the door schedules. Now, the door schedules, because they were supplied as part of four different packages, the door schedules were all completely different. And they weren’t provided as spreadsheets, they were just images in an O&M PDF. So for us to do anything with them we’ve had to strip them out of the PDF and convert them back to data which obviously in itself is a dodgy thing to do, but that’s what we’ve done. And then we can start to find where the inconsistencies are.

But all of that is unnecessary if we provide the information in a machine-readable form. And the final thing is that a lot of the things that we need from an asset management point of view are often never delivered in the BIM models because it’s the level of detail that, for example, designers don’t go down to because it’s not really part of your scope, in many cases.

JAMES DIXON Do you have an example of that, George?

GEORGE STEVENSON Yeah, I would say the actuators are very seldom modelled. Door closers aren’t normally modelled. Now obviously they should be in schedules which are referenced properly to the door type. But we also find that door types aren’t provided either, we end up just with instances of doors. And this is competence of the people that are creating models in many cases, but also I think to some extent the fragmentation.

PIERRE SAUNAL Just on that, maybe more of a question than an answer, is it brings the question of how much we model and how much we input in terms of data. Because you mentioned door stoppers and a lot of element. Our view is they shouldn’t be modelled because they create sometime unnecessary additional geometry which make the models heavier. However, the data should be embedded in it at the right location to be found by the right people. So just that big question of frugality in terms of modelling is always an interesting one.

GEORGE STEVENSON It is, but I would also say that one of the challenges with that is if we’re reliant on IFC as the mechanism of exchanging that information then the IFC definitions often don’t have adequate detail to be able to do it. So you can have situations where things are lost because we’re using, understandably, the open standard to exchange things.

PIERRE VENTER I think the thing that’s really important, my approach to this thing is what I call the placeholder. So actually a door, as long as the geometry and the spacial parts are correct, in other words a door is a door set, it’s got frame sets, ironmongery, lock and latch sets etc. Is that in fact those pieces could be separated. The geometry, the spatial issues, as long as that’s correct it’s going to be in the right place. But all of the components that go into that could in fact be a list, if you were talking about Dynamo or any of the data set processes, And what it is the two tie together, and George that goes in into the IFC.

But what it is is the term that I’m going to use is called the performance spec. In other words if I have a door and it’s a fire door it says it’s got to perform in a certain way, it’s got to operate in a certain way. And that can actually be done in a digitally readable form because now you get the geometry and you actually get the data at the same time.Yes, this is a text file, but you can actually go that little bit further. And George, you could be putting the Uniclass classification in there as well. So now not only are you picking ip that there’s got to be a door closer, but you know exactly what category it belongs to, what classification it belongs to which starts tying it up further down. So, it’s a little bit simplistic, but that’s the one mechanism.

JAMES DIXON Yeah, an architect needs to produce a door schedule that tallies with what components relate to that door. That’s not BIM, that’s not digital, that for time immemorial is something they need to produce. So parametrically having a basic description, a little bit of text that identifies this as an actuator or a door closer enables the architect to produce the schedule they always needed to produce.

Once it’s parametrically in there we can export that with P sets into the IFC reliably, consistently, every single time and produce your COBie. It is linear. I’ve seen it, I completely agree, but you don’t need to model for model’s sake. A door closer, you could probably live without, but it does need to be identified against the right doors. And that has to be in the schedule. They do it parametrically, their schedule is just done automatically. It’s not complicated and it’s not onerous, in my opinion.

GEORGE STEVENSON And when it’s handed over it’s got to be machine-readable rather than just in a PDF.

JAMES DIXON Totally, COBie, exactly, that’s why I personally like COBie because it is consistent every single time. I’ve said it in front of people in conferences and there is an audible gasp, but it’s standardisation, very straightforward. Go to this location for this data, easy. And you can teach a CAFM that very easily, you don’t need anything sophisticated to teach CAFM to read cells or columns or rows.

ANDREW DE SILVA One thing I’ll just share which is different to what I was just about to share actually which is something that we’ve been working within the office. We’re creating content with the aim of making sure that it’s consistent and compliant and aligned to the specification, all digitally tagged. This is the kind of view into, for example, the internal walls but within our content library we’ve created not just things like walls, ceilings etc, but also looking at things in terms of assemblies. So that could be to do with, for example, bathrooms which are compliant in terms of their geometrical relationship to different parts within the sink and the bath and the WC etc.

So I think this is picking up on some of the things that’s just been talked about where the aim is that Andy from the team has created a plugin into Revit which means that all of our info is held in Notion here. But it means that within Revit the team members, we’re kind of also changing the way that the team is structured, but the idea is that you can go into this content library that we’ve created. Drill down into what is required from that component or assembly and then choose the relevant one. And when you bring it in it’s got the information, but it’s also got the geometry and it’s also got the things tagged and specified, linking it to our master specification already.

So that’s something we’ve tried to do and again, it’s more powerful also when we’ve got the assemblies because there are so many permutations of an accessible in a Part M43 bathroom. And you would waste a lot of time if you’re asking junior members to read Part M, try to unpick the requirements and then design it into one of the schemes. So we’re trying to get a large amount of whatever the sector or the project, a large amount of it is pre authored and created which can then be brought in. Clearly there is going to be bespoke elements because all projects will have bespoke elements and that’s where architectural interest or the site specific requirements kick in.

And then Andy and Lee who is the head of spec and the R&D team are working on it to make sure that it is given a particular name, so then it does allow a door to be aligned tot the ironmongery set, to the specification etc. So that’s something we’ve been doing to try and standardise the components and make sure that they’re all working or compliant. Which helps us as a practice from the Building Safety Act but also just generally meeting the employer’s requirements or the regulations. That’s something we’ve been working on for a long time.

JAMES DIXON I love that. One quick observation, if you have the BG white book product data reference in there, because that is the one thing that commercials will use to talk back.

ANDREW DE SILVA We do have that in there as well, that’s embedded. In terms of our little ecosystem of software that we’re using, we’ve got Revit, we’ve got MBS in terms of the specification side of things and then Notion is what you saw. But Notion is kind of like an intranet for all sorts of things within the practice, all centred around the idea of storing information in database format so that it can be machine-readable, it can be filtered, it can talk to each other as such. So that’s something that Andy in our team who is an architect but historically is also leading on the R&D side, so he’s very much aligned to that database mentality.

We’re trying in information in the master specification and the content library, things to do with obviously carbon. So taken from the databases I guess which are available in terms of different materials. So we can extract onto these and get high level guidance for the projects. Clearly there will be more specialists so we can drill down into more detail, but just to give us an idea of initial moves that we might make. So that’s something that we’re building in as well. We’ve still got a fair amount to go in terms of picking up all of the components that form various building types. We’ve started predominantly with internals, floors, ceilings, doors, bathrooms, kitchens, doing the envelope at the moment.

So like I said, most scenarios we probably have something that’s the correct buildup specified etc. So we’re testing it out on a school at the moment. And then on this scheme, in one way it’s very simple the task. This is the resi scheme built 25 years ago. Surprisingly or unsurprisingly, there is no real information about what was built, even though it’s only 25 years old. No as-builts, tried to contact people like the NHBC who were the approved inspectors at the time, no information. So really having to start from scratch which is unfortunately not that uncommon in terms of what we’re trying to do generally.

It’s treated as a pilot, the good thing is that most of the building is, from a combustible point of view, is brick. So actually we’re only talking about specific areas which is related to this timber cladding and the timber decking on the balconies. So again, from a first pilot project that the team are going through it’s a really good one to do. But it just shows you still the complexities there are. Waites had a point cloud survey conducted for the entire estate. So Pierre, picking up on some of the things that you do. That surveyor then produced a Revit model. As you can see, fairly basic in terms of internal elements, but externally accurate geometrically which is the key thing.

And from that we’ve then generated this document called a sketchbook, but essentially it’s looking at what are the different details that we need to resolve. So one thing is that timber cladding facade. We’ve tried to look at what’s existing, keeping in mind that we don’t have any as-built information.

JAMES DIXON Is that intrusive? How do you know the wall buildup with the point cloud?

ANDREW DE SILVA Good question, I’ll come on to that. We've made some assumptions in here, so all of the red text is assumptions. Before we started a company called OFR who are fire consultants, they did a fire risk assessment external walls review which included some intrusive surveys onsite, externally. So for example, that is why we know what the buildup of this timber cladding is, they cut holes in it. So some things we were able to know because they have been physically opened up. And then we were then generating proposed buildups, again based on a fair amount of assumptions in terms of the text here.

So that’s what we started off looking at and you can see from these photographs. These were done by the initial fire risk assessment external walls. And that’s enabled us to take the point cloud survey which had generic wall thicknesses, 250, 280 mil thick. We then broke that model to create a more granular version of the external wall based on what we’d found. It still didn’t give us the full picture, there were still areas which we didn’t know. Predominantly what is behind this breathing membrane and how thick is the blockwork etc behind that. That was basically where we started. So you can see we’ve looked at those kind of things and started to generate the existing and the proposed for all of the different areas within that curved timber.

We also had some instances where there were like an interface between the brickwork and render. So the areas we were looking to do were essentially based on the fire risk assessment and areas that had to be remediated in order to make the risk tolerable. And the timber decking, the balconies all had timber decking, so that had to be removed, And finally there were areas where they found EPS insulation. Most of the scheme already had mineral, so that was good, but there were some areas, these projected bay windows that had EPS insulation. So that’s something where we understood some elements of it, but not the whole picture.

This is the views from the 3D survey model. And we were kind of identifying where they could be risks in terms of other EPS insulation etc that we may find. So what was good here is that even though the building was built only 25 years ago, from a data and information point of view we had very little. The fire risk assessment intrusive survey was a good start. Then commissioning the point cloud survey to generate accurate information about the as-built status was also very good. And we asked them to do the internal version of that as well because there were some things that we had to do internally.

JAMES DIXON When the laser scan, how well did that facade buildup actually match the O&M? I'm just interested, did the O&M actually match how they built it in terms of the external envelope?

ANDREW DE SILVA There was no O&M that we could find. And so what we were looking at is them doing further surveys.

SYPRIAN A PITKIN What I’ve seen, it’s extremely good. However, you can’t say something or keep using with regards to assumptions. I’m coming from our design and compliance consultant and I do advise BRPD and I do the role myself. From experience, especially with learning projects, we are finding it has been advised when you reach out to the BSR is that they understand a lot of these properties, these buildings are occupied, but to avoid the risk of doing the project to be stopping to apply for change control is advised that we get as much existing information.

So it’s advised that you do 100%. If you know that you’re doing a buildup and there are other elements that you’re unsure of it can affect the application. And it’s advised that you get as much information and avoid the assumptions because from experience applications have been turned because of that. One of our applications went through simply because it was a hospital and you can’t do certain works in an occupied hospital. So we had to do a statement to say look, this is what will we do because of the occupied hospital we can’t do certain intrusive work.

However for the block work, for the facade projects that we have done we are getting RFIs asking we need more information. You are making that assumption, we notice you’ve not been giving information on what material has existed there, we need to know that whatever your putting in they actually integrate and it complies and there’s test evidence that you won’t have a negative reaction with the material. So just putting in there that we need to avoid the assumptions at this stage.

ANDREW DE SILVA I 100% agree and what we did is, this document was based on in terms of what information we knew from the initial fire risk assessment external wall survey. And then what we were doing in this document is highlighting in all of these details with the different red and orange and blue text, which items that we didn’t know about before we then essentially went back to site to do a further set of intrusive surveys because we had too many unknowns. What we then did was I’ll show you some of these. We had generated that report from the information that we had at the moment, we then commissioned, this is both with structural engineer, ourselves, the fire consultants, essentially a markup of those elevations to identify additional intrusive surveys we wanted.

To pick up exactly that point you made about the unknowns. For example, what we did was because we weren’t 100% sure, we knew what the timber cladding buildup to the blockwork was, but we didn’t know beyond that. So for example, we then went back and had a large area removed over the two different areas in the two different blocks as well. Again, to pick up on your point, Syprian, about not making an assumption about just because it was built one way in one block that it wasn’t built the same way on the other block. So we opened up both of those blocks and what was good is that we found the same buildup i.e. essentially concrete slab, blockwork, steel posts and then the timber cladding.

There is no insulation on either of those. What we didn’t know was the thickness of the blockwork because, coming back to your point James, in the survey model was suggesting a 280 mm buildup from outside of the timber to the internal space of apartments. And the numbers don’t add up because we had something like 50 or 100 mil timber building, then that would leave somewhere in the region of 180 mil, but what was 180 mil based on? Was it based on the blockwork laid on its side? Which is something that was fairly common practice back in the late 90s. Or was it two layers of blockwork or something else? So we had a core test done which sort of helped because the core drill was 150 mil thick.

It showed that there was clearly one layer of blockwork, then a membrane and then something else beyond it. Again, when you did the numbers it wasn’t quite adding up in terms of the survey model. So since this intrusive survey which as you can see we went through different parts of the different blocks. So, have three different blocks that we tested out all of the surveys on the opening up works. We’re going to go back again to drill from the inside out because there are still unknowns. And we’re building that into our digital model from the survey that was done. so trying to fill the gaps and each time we do some more intrusive surveys we’re picking up on something else that we need to know.

To pick up that point, Syprian, that you made, I want to get to a stage where in our details and the specifications stuff we don’t have any red notes where we don’t know what that buildup is. So all ambiguity has been taken out by carrying our surveys. The structural engineer is now doing the calculations in terms of fixing to the slab edge which is probably the most realistic solution in terms of the new rain screen cladding system. Just because we don’t know that we can see from the surveys that the blockwork doesn’t look like it’s tied in very well to the slabs etc. So we’re kind of bypassing that because it’s too much unknowns and then going on to fixing to the slab edge, for example. So we’ve got to engage the specialist subcontractors which will happen in the next couple of weeks.

And that’s to do with the rain screen, there’s injector clad and the detectors and the balcony replacements and stuff. So all of the specialist subcontractors will get appointed soon and then we will develop the final construction details specifications, data sheets etc, to then go to the Gateway submission at the end of January is the target.

PIERRE VENTER Just a question, something that I’m busy doing at the moment which is looking at the scan from the outside. And in those areas that you’ve identified at the earliest possible time to scan from the inside, so you’re actually getting the thickness of the buildup. And in the same common areas across two buildings, before you expose them. And it’s actually been quite useful to have a look at what you anticipate being there and pretty much what did appear.

But the bigger part for me would be a record for buildings in this particular location, this area, this region where typical construction methods might be pretty consistent across the board at different levels: Tier 1, Tier 2 and white van brigade. But in reality, just to give you some indicators to how things actually happened. And that would sort of be not an assumption so much as simply opening a common practice and an understanding. And that then becomes part of a guideline to say consider this.

ANDREW DE SILVA That’s right Pierre, because again what we found is before we went on to site, the original fire risk assessment had opened up various things. But the weird thing was one of the bays EPS was used as the insulation. But we were like well, was that material used in other bay windows as well? So, as part of the second intrusive survey core drills were taken into all of these spandrels to make sure that there wasn’t any other areas that a bit of rogue EPS was thrown in by the contractor. Luckily for us there wasn’t any other found, it was just in that one bay that they had decided to use EPS rather than mineral wool, for example.

And that’s also why we wanted to check across the two buildings to make sure that the timber cladding buildup was the same across the both as well. Because clearly one might have been built slightly later, there might have been some material supply issues and they couldn’t get everything in time, so they changed the material. So that’s where the danger lies and the need for quite expansive surveys to make sure that you haven’t missed something. Which then could be a problem when you get to site in terms of the change control thing, Syprian that you picked up and noted as well.

So, we’ve still got a fair amount to go and as you can imagine it’s sort of now, there’s elements internally that’s also coming up in terms of there was some maintenance stuff that hadn’t clearly happened. So was it built properly in the first place? Or has it simply not been maintained properly? So they’re getting into grey areas at the moment, but nevertheless on terms of facade we are yet needing to take one more round of intrusive surveys and then feeding that into the structural engineer, the specialist subcontractors that will get appointed and the facade consultant to make sure that what we’re putting back s compliant and is suitable for this particular building type. So that’s where we are in terms of that process.

JMES DIXON You mentioned material shortage. Were they two different buildings? The one at the front and the one at the back?…two different subcontractors…I appreciate it’s a remediation, but what it does reinforce is if the detail was to not have that membrane then how is it possible that the building was actually constructed in some locations with that membrane? If you have a proper pro regime where you can evidence how you built something, we would try something we've gone in not quite clearly a brief detail, just that we clearly know the way one facade that's identical to the next in a different direction has been constructed differently.

ANDREW DE SILVA Yeah, who knows what happened in that particular, because if it was using EPS on all of the building where that particular detail is then that would sort of make sense that maybe a different party were using it. But it’s only in one bay of the entire two buildings. So, we could kind of go round and round trying to figure it out, but the bottom line if you’re doing the project now, things like the technologies in terms of scanning, the 360 photography, the time lapse, all of those things are what we’re learning from this project what not to do in terms of record keeping.

I guess that’s what you’ve be doing as main contractor to be able to evidence and strip back the times. A bit like Google maps and its archive photographs so you can move the toggle and see what it was 20 years ago and what it is now.

JAMES DIXON It’s competency, you’re the subcontractor, how will you make sure that you are building to the approved detail and that is a QA matrix, that is a QA plan, we will do this, we will evidence it this way. Yeah, that is a great example of non-standardisation on one project alone.

ANDREW DE SILVA The other thing I want to show you, there’s quite a helpful webinar by the HSE that just came out, I’ll share the link. There’s a gentleman on this video which basically goes through the online portal showing what information is required to make a submission. And whether it’s Category A works, Category B, Gateway 2 and Gateway 3. We had an initial Gateway checklist and a Gateway 3 checklist and my colleague Lee has been looking at this video and just cross referencing to see if we have missed anything. Because obviously the Building Safety Regulator has been changing and I guess improving the portal and the way that we can upload drawings etc.

So clearly it’s going to evolve, but this is quite a good guide and I’ll share it on the chat for everyone to be able to watch that. It is good to dry run it virtually. That’s something that we’re using just to make sure that we’re thinking forward to Gateway 3, Pierre, you’re experience will be really important. Because i think a lot of people might get focused on Gateway 2 for Gateway 3 and actually we need to be focusing on the handover and getting the certification and work back in case we’ve missed something that is too late to pick up later down the line.

GEORGE STEVENSON I think that’s a really good point. We looked at doing this probably three years ago when people were registering buildings. And one of the challenges was that people were registering the building and when they went on to register they didn’t have all of the information that they needed to be able to do it, but they didn’t know what questions were going to be asked. And there seemed to be a reluctance from the Regulator to provide a schema that would determine what information was needed. So, from what you’ve seen there Das, do you think we’d be able to generate a standardised schema for the information to be collected and then submitted?

ANDREW DE SILVA Yes, I would say so because it goes through each of the questions and it also shows the information types drawings, compliance or statements etc. So I think it’s a really good guide. Obviously it doesn’t show you what needs to be in a compliance statement, but that’s up to the team to agree depending on the scale of the project etc. But that’s what we’re doing, just updating our Gateway 2 and Gateway 3 checklist based on this video. But I think it was very clear. Things will evolve, but it’s a really good overview I think. Starting an application online you don’t know what the questions are, so this person from the HSE is basically doing a dummy run through the whole thing.

GEORGE STEVENSON So is Lee, in that exercise he’s going through, is he actually determining what those questions are?

ANDREW DE SILVA Well, this video is showing the entire process, so we’re looking at the video that they’ve done, just taking the questions and the sub-questions for each of the bit of topics and creating our database checklist. And then we’ll share that with the team and say party A needs to supply this, party B needs to supply this, the client needs to supply that. So that’s what we’re doing at the moment. Once we do it I’m more than happy to share it because it’s nothing secret.

GEORGE STEVENSON I’m just wondering whether we can break the job down so that you’re not doing it all.

ANDREW DE SILVA I think that Lee has actually done it quite quickly already. And these will obviously help create a database back of house kind of. We want to present it in a Word way, but it’s been generated as a database in the background

GEORGE STEVENSON Brilliant, well done, that looks excellent.

ANDREW DE SILVA I just came across the video, someone had shared it. It is really useful just to be able to dry run because until you do you don’t know. And it allows us to really manage what the team needs to produce. I wanted to share that, I realise that some of it may not be exactly relevant to the audience, but I think in terms of data, having the right information, going through the right processes, having the right consultants. And having a realistic programme which to be fair on this project Waites are getting input from everyone, has made it really quite a good pilot for us all to be involved in.

Because it will be rolled out across many projects and I’m sure everyone in this team will be working on projects where similar things are required, whether new or refurb. We’ll aim to submit at the end of January and we’ll keep everyone updated. In the next sessions like this what I try to do is to share the next update, so that could be some of the trackers that I talked about. But it’s also really nice to get input from your side, so anyone on the call, because this is just my experience and other people have different.

So Pierre, from your Gateway 3 submission, it would be great to hear more of that if that’s possible. So what are the things that we should be looking at on the project that I’m working on or the wider team. I don’t know whether I could volunteer you for a little show and tell next time?

PIERRE SAUNAL Yes, that would be my pleasure, I’ll try to be as clear as what you’ve shown. It’s very clear, and I like the approach and I think what we experienced with having to go through Gateway 3 was actually a live check of our workflows. Because we had to retrospectively basically self assess how we developed the design for construction to make sure that no stone was left unturned. I think that was really the key thing because we had behind us funders, insurers, obviously Building Control looking at absolutely everything. So we had to model everything on the project and make sure more than just the geometry that the data was there. We had to be sure that every input within the model from a design perspective or coordination perspective, there was a reason behind it.

I think that’s really the key thing. You can’t have any input in your data without knowing why, what you are putting in. Because you are going to be questioned, you are going to explain why. And it’s almost better to know that some data is not in it and why is that, than just working or progressing blindly in developing your model.

ANDREW DE SILVA We can arrange a date with Pierre, it will probably be late October or early November. Because not many people have gone through the Gateway 3 process, more and more people are great news going through the Gateway 2 successfully, it will be really valuable for everyone.

PIERRE SAUNAL I’m very happy to be very open and transparent on the documents that we had to prepare as well. It’s one thing to have the BIM model, but then you need to format the data in a number of documents, or more the old fashioned way which can be fully extracted from the BIM model. Because it’s not just about the design and the model itself, it’s about the process. How did we get the right information at the right time? Did we question it? How did we federate it, coordinate it? So, it’s a sort of hybrid old fashioned approach and more modern with the BIM model itself.

JAMES DIXON So that conversation earlier about what to model and what not to model, if it doesn’t offer anything from spatial coordination perspective that is probably a good conversation point to have. After that, the data, you can have a very basic simple model but with lots of data that sits behind it. So the door and the door stopper, I’d probably stop at the door and then embed the data to say that that is a fire door with a stopper that sits with it. That’s how I would approach it because Pierre, you just mentioned that everything had to be modelled. Was that to facilitate the data export? Or was it something else that made them want everything to be modelled?

PIERRE SAUNAL First I should have said we work mainly in Archicad, that’s our main authoring tool. We use Revit because we developed hybrid workflows between the two for a very large project in Saudi, now in Nigeria because there is still a demand in BIM protocols to use Revit. We know that it’s some time against the ISO, but it’s to be realistic based on the country that you work in. For example, we’re bidding in Senegal for a very large campus, a public project, and in Senegal the BIM environment is built around Revit. Whether we like it or not you have to deal with it.

But our main authoring tool is Archicad which allows for some capabilities in modelling so we can have heavier models, more details. So we try to model as much as we can, depending on the stage, to have the information linked in 2D, 3D and data. So we extract all of the information from the BIM model. We don’t do any 2D, it’s on the input of geometry and data. But that being said, we try not to model too much. I really like what you said, James, about a door. You need to have realistic representation of the door. However, you probably don’t need to model all the components of the door.

You can embed a lot of data within he BIM objects. So it’s always to find that balance, depending on the output which is required by the client because we didn’t talk about it, but it goes without saying that the fee is proportionate to the amount of geometry and data you embed in your model. And sometime that is a contradiction. It is difficult to convince a client that you need to agree a certain amount of fee to be able to embed a certain amount of geometry and data in your model, based on the requirements, for example, for Gateway 3. So, it’s a very interesting discussion on that topic and unfortunately I don’t think there is one answer and that is the problem. We try to have a generic workflow which is still very project specific.

JAMES DIXON One point I’d make is COBie and how it didn’t reflect the BIM model. I’ve actually had a client say to me don’t worry, you don’t need to generate this data from the model. To which I would say there’s absolutely no way I’m going through 250,000 lines of COBie Excel spreadsheets to populate it. That would make my life harder. So I’ve had that conversation with clients and i’ve said no, we will be doing it in the model because it’s just safer and easier actually to get there.

I would add that into a client’s competency. We’ve spoken about BSA, how are you producing your door schedules? Take me through it as an example. And if there is a well, we will produce the GM and then we will update and we. That for me there is a competency problem there and they will fall over and make mistakes.

GEORGE STEVENSON Just to make you aware of the work that was done last week that you’ve got access to. For several years we’ve been building datasets with subject matter experts on those asset types. So we've got detailed information that probably upwards of 100 people who are subject matter experts in fire doors, penetration seals and things like that have gone through and done a deep dive explaining what information they need. I'll just show you one example of what that looks like. So, for a fire door for example. This was originally done in 2021, it was last reviewed in 23 and we reviewed them again last week.

So we had a team of people looking at these and it breaks down all of the different elements of what goes into a door and what that looks like. But also it starts from the point of view of the Building Safety Act and what the Regulator is looking for. What risks is that particular asset protecting against? What do people do to the asset to stop it from working? So this is very valuable stuff, it’s come from people on the ground. Also what information do you need from a point of view of specification, what materials are being used? The construction detail, the installation detail, inspection and maintenance. What tasks are needed? And what competency is needed? And how should change be managed?

So, we’ve gone through that exercise with all of those different asset types and last week we ran the workshops, there all recorded. But what we’ve also done which is new is we’ve generated an AI podcast that’s been generated from the conversation. And I have to say I think it’s extraordinary and if you’re interested I’ll send it to you. The reason I’m saying all of this is this is also part of the whole process. And what we also did, we broke the asset types into groupings. So that’s compartmentation, that’s largely smoke control. Here we’ve got sprinkler systems and risers and here we’ve got emergency lighting and escape signage. I have to say that we didn’t get onto that because we spent all of our time on signage and the like.

We’ve also got a group starting in a couple of weeks’ time on lifts, this is a brand new one. We’re going to have a group of people who are subject matter experts on lifts telling us what information is needed.

JAMES DIXON That will be fantastic, because working with lift manufacturers in this context is extraordinarily difficult. So I just say, yes, they don't want to play at all.

GEORGE STEVENSON Yeah, we seem to be breaking through that, James. If you want to be part of any of those, the first thing is to be able to use and access what we’ve already produced. What we’re also doing from that exercise, we’re going to spin things off because we want to start creating AI agents that can then interrogate that information and then identify what’s needed and what isn’t needed. Because the challenging thing that I’m coming across all fo the time is the interrelationship of this information that is so critical.

ANDREW DE SILVA That brought nicely into the M&E world for Franco, some thoughts or reflections or any observations.

FRANCO PITTONI I would like to make some general considerations. What was said in this meeting is very much related to the concept of level of detail or level of definition. As far as I know, there is no rule, standard, regulation about what should be modelled. We've talked about common sense, but as far as I know, there is no clear definition of where the modelling effort should stop. This is the first consideration, and this is in my opinion, a very, very important aspect because it leaves this decision to whoever is modelling. The second aspect is if we agree that something should not be modelled and therefore all the data should sit somewhere else.

For example, the elements of a door, the information about the elements of a door should be contained somewhere else. I think that there is another challenge here, which is how should we structure the level of information of the data sitting outside of the model? For example, looking at the M&E world, if you open an electrical panel, as we all know you will see many, many small things, right? Each one of them may have a really important function in how the building is performing or functioning. So for example, if we decide that the panel should be only be modelled in terms of its geometry and what is inside should not be modelled, for example, then we need to decide how we organise the information of the mini circuit, breakers and switches, whatever is contained in the panel.

SoI think that these two topics, where is this line between what should be modelled and should not be modelled? And the second topic is how should we organise the information of the elements that we do not model? I think these two questions are really important from my experience and point of view.

GEORGE STEVENSON Just on that point, one of the things also that is becoming apparent, let’s say Revit, Pierre, I understand the need for not to just be proprietary. What we’re finding is if a project is going over maybe 3 years it’s a real problem because when the model is handed over it’s probably still in 2022. And unless it’s been upgraded progressively by everybody, and a lot of people don’t do that, then what you’ve then got is a problem. And we’ve just had it recently on a project where we’ve had to upgrade from all of the models, and there’s a lot of data, a lot of models, and we’ve had to upgrade it year by year. This is retrospectively.

And it’s fortunate that we have all of the versions of the software. But a new person coming along wouldn’t be able to do that. And also when you go from one version to another you can very easily experience data loss or corruption. So I’m just flagging that up because I don’t think it’s something that is often put into EIRs that the version that is handed over should be the current version. It shouldn’t be down to the client or the asset manager to figure out how they’re going to upgrade files that are years old.

FRANCO PITTONI There is another aspect which George knows about which is my passion and expertise in testing and commissioning and handover. I can say with some decades of experience behind me that very, very rarely as-built data and commissioning data are handed over in a proper manner. Which makes the asset owner suffer from the lack of correctness of what he sees or how things should function. So this is another very important aspect, to have an as-built dataset, not only the geometrical one but also the data one.

ANDREW DE SILVA Thank you and I’m sure there’s lots of knowledge about all of that in the team which we can continue to talk about.

PIERRE VENTER With Franco’s point there about what do you model to do the DB and everything else separately. I think there’s an important point one to actually have a look at what I call scenario planning. In other words, when you are actually going to be in the field you would typically need a circuit diagram to say where are the breakers and where’s the circuit. So in other words, the floor plans should indicate that because I’ve had that onsite in a Mall previously where the circuits because the tenancy is changing and all of the rest of the dynamics that George was talking about, actually had an impact on the ground.

The socket outlet wasn’t connected to that DB, the circuit would have been correct in the model, but it was elsewhere.The other part I think that’s really important is the commissioning part. That the content that is as-constructed, I’m going to use the term rather than as-built, is the records are never kept up-to-date and in particular with the bigger contractors where they’ve got subcontractors, the subcontractors do not have the competence and the capacity to deliver models of a decent standard. So i think those are all thing9s that we really do need to put in as important derivatives.

George, finally on the story about the versions, we are moving in to the open world and one of these days Autodesk will have woken up from their slumber and changed that. But I work for a big practice in London and we were on version 2020 and they were still reviewing and updating models of existing facilities on 2014. Fortunately Autodesk keep all of those copies available, but it is an absolute ache having to go year by year and do the quality checking was an enormous resource sponge. And who pays for it is that big problem.

PIERRE SAUNAL That idea about software version is coming to an end, whatever software you are using. Because the end product in terms of as-built model has to be IFC, always had to be. Different people in the design team can produce a proprietary file if needed and sometimes its included in the BIM protocol. But the reason why we believe it’s over and it is things to what you said, Pierre, about the capabilities of main contractors. We re 99% there, to give you a practical example, on the new stadium in Casablanca, the Hassan 2 115,000.

We are in final discussion with the two co-contractors that won it which are the biggest one in Morocco. The reason why we are there is not because we are a big practice, it’s because we are bringing what we call the hybrid workflow where we’re going to use Archicad for the master model to be able to have in one file the entire stadium. And Revit for what we call the micro models because we have to work with different disciplines. Combining all of that means it’s impossible to give an as-built model in any proprietary format because the information, geometry and data is scattered between this meta model and micro models. The only way to have an as-built model which encompasses everything is IFC.

And that is the only way to give an as-built model to the end client. So I think that discussion about what version we use and upgrading is still true, but I think it’s slowly and thankfully, from my perspective, getting out of the picture.

GEORGE STEVENSON I’m intrigued as to how you deal with parametrics in IFC, Pierre.

PIERRE SAUNAL Between the two file formats, you mean?

GEORGE STEVENSON Well, Revit, for example, IFC is like PDF.

PIERRE SAUNAL Yes, you can get a lot of parametricity between one and another and back and forth, so…

GEORGE STEVENSON I don’t think so. Maybe I’m wrong, the last conversation I had on this was Thomas Liebisch who is one of the authors of IFC4. And he said that it didn’t support parametrics.

PIERRE SAUNAL To give you a practical example with SNAP twins, I shared the link, our algorithm creates the IFC space automatically from the point cloud. And these IFC spaces, you import them in Archicad, they become a native parametric zone with all of the data and the parametricity. You import them in Revit and they become native rooms with all of the data and the parametricity. So it is possible to move.

GEORGE STEVENSON Yeah, that’s fairly basic. What I mean by parametric is where you’ve got a door and a wall, so you place the door in the wall and it Revit there is a series of parametrics that actually go to make that work.

PIERRE SAUNAL I’m going to bring Jaime who is really the brain in BIM. That’s where the interoperability between the software is important and that’s why you think Archicad is another strength because NEMETSCHEK, owning Graphisoft and developing Archicad, is insisting a lot on the interoperability between Revit and Archicad in that instance, but there are other solutions. Which with our workflow allows anyone to use Revit to model and to a very high degree we can bring all of that in Archicad, for example, and keep the parametricity.

The challenge which has been since we started that hybrid workflow in 2001 is more the extraction of 2D information because Archicad and Revit are fundamentally different. So whatever you create in Revit you can bring it in Archicad, keep the parametricity, but you don’t keep the 2D output. So there are a number of tricks that you can use to keep that, but I would say the technological limitation at the moment is to order 2D output between the two back and forth. We are working on it, but that will never be solved because Autodesk and NEMETSCHEK are happy to bring the interoperability to a certain level, but will never develop a unified solution unfortunately.

ANDREW DE SILVA Thank you, Pierre, maybe you can touch on some of those things in the presentation, and I’ll email you about a time. Great to meet lots of new people today, we’ll see you next time.

CHAT

Joe Stott (External)… Shouldn't we be looking at how the specification should be made machine readable? Spec always takes precedent over drawings / models contractually.

P T Andrew de Silva (External)… Walk through how to use the digital portal to manage a building control application

pierre@airc.digital (Unverified)… Thank you all for a very interesting and informative discussion. For those who want to have a look, here is the link to our automated Scan-to-BIM solution https://snaptwin.eu and we look forward to receiving your feedback (and .e57 if possible )And my email to follow-up is pierre@airc.digital

Academic Summary - Design Working Group Meeting BIM4Housing (30-09-2025)

Purpose & Context

• The Design Working Group meeting focused on exploring the integration of BIM methodologies and design processes across real projects. The session featured architects and design leads reflecting on challenges, coordination methods, and project-specific lessons from BIM implementation.
• Discussion emphasised practical design considerations, data integrity, and the balance between digital modelling accuracy and real-world constraints.

1. Integration of BIM into Design Practice

• Participants examined how BIM supports collaboration across design disciplines and its limitations in current project workflows.
• Several architects described BIM as a coordination and communication tool that is still not fully leveraged for decision-making. One commented: “We often use it more as a drawing tool than a design intelligence tool.”
• Emphasis was placed on embedding performance-based outcomes in the design model rather than post-design compliance checks.

2. Coordination Across Design Teams

• Participants discussed design team alignment, focusing on federated models and clash detection workflows.
• One participant noted: “Clash detection identifies conflicts, but it doesn’t necessarily solve them – you still need human judgment.”
• There was agreement that effective coordination depends on early engagement and a shared understanding of design intent, not just technology integration.

3. Challenges of Data Accuracy and Interoperability

• Architects raised ongoing issues with differing software platforms and levels of model detail (LOD).
• One architect stated: “We spend too much time trying to reconcile models rather than focusing on the actual design.”
• Participants agreed that interoperability standards (e.g., IFC, ISO 19650) help, but the problem often lies in inconsistent execution rather than the standard itself.

4. Design Intent vs. Construction Reality

• Several contributors reflected on how the purity of digital design models is compromised during construction. “You design a perfect model, but site conditions will always introduce imperfection.”
• Discussions covered the need for better feedback loops between site and design offices to capture real-world modifications in the model.
• Some suggested structured ‘lessons learned’ sessions post-handover to feed into early-stage design on future projects.

5. Managing Information Flow and Version Control

• The group discussed the difficulty of managing revisions across consultants, particularly where multiple design updates occur simultaneously.
• One BIM coordinator noted: “Everyone assumes their model is the latest one – until you find three versions of the same file in the shared folder.”
• Version tracking and controlled model exchange protocols were identified as critical, with recommendations for stricter information management planning under ISO 19650 principles.

6. Digital Competence and Cultural Barriers

• Participants acknowledged that varying digital competencies within project teams create inconsistencies.
• One contributor stated: “Some architects are digitally fluent; others still prefer paper sketches and PDFs – the culture shift is uneven.”
• The group discussed the need for upskilling and embedding digital thinking as part of architectural education and professional development.

7. Reflections on Design Quality and Responsibility

• The conversation turned toward the ethical and qualitative dimensions of design in a data-driven environment.
• An architect reflected: “Technology can streamline delivery, but design quality comes from creative judgment, not software automation.”
• There was consensus that digital design should enable, not constrain, architectural creativity, and that responsibility for design integrity lies with individuals, not systems.

Emerging Themes

• Persistent gap between BIM’s technical capabilities and its use as a creative and coordination tool.
• Need for structured design feedback and learning loops post-handover.
• Interoperability and version control remain ongoing challenges across multi-disciplinary teams.
• Cultural and competence differences within design teams continue to hinder digital integration.
• Key Insights & Lessons Learned - Design Working Group Meeting BIM4Housing (30-09-2025)