BIM IN ASSET MANAGEMENT MONTHLY FORUM 14-01-2026
GEORGE STEVENSON (shares screen). For those that are new, we started 5 years ago with BIM4Housing. And about 2 years ago we started BIM In Asset Management which is looking specifically at the asset management side of things. And the approach that we took with BIM4Housing was to focus on better information management and look at BIM from the point of view of what information could the development teams produce that would make the job easier for design and construction. And what information could design and construction teams provide that’s going to make the job easier for the operations teams.
And as a result of that we then looked at that from a point of view of the golden thread to say we needed to improve the way data was structured. So we came up with standardised information requirements and standardised data templates. We also then looked at processes to make sure that you could improve the way that that information could be provided. And that’s where things like classifications and reusable libraries come in. We did a lot of work specifically around fire safety because that was part of the process that the government asked us to look at. Sustainability was also an important aspect and the production of digital records.
The way in which we went about that, which we’ve learnt a huge amount as to how you can get knowledge from people and then turn it into something meaningful. We’re part of what’s called the Golden Thread Initiative which was the group that was advising the government and the regulators in terms of what was reasonable to expect the industry to provide. So we became part of an industry initiative and our focus was on the asset and survey information. And the way in which we went about doing that was to create Teams meetings and some of them in person where we’ve got groups of people from a particular discipline…manufacturers, designers and operators.
And the beauty of this is we’ve got people who are probably competitors sitting around the table sharing their experiences. And what we were really addressing was the fact that the legislation doesn’t give any explicit requirements. And therefore you’ve got different experts interpreting the information slightly differently. And they are then giving guidance to people which is not consistent, so that confused the heck out of supply chains. So what we’ve been doing is trying to create where we can reusable libraries and standardised processes.
And this, for example, is a Miro board that we created to get everybody who was interpreting the legislation to do so in a collaborative way. And this is available to you, it’s a Miro board and if you want access to it then put ‘Miro’ in the chat. The other thing that we were doing, these are publications that you can get off of the BIM4Housing website. We’re looking at what risk does an asset mitigate? What do people do to it to stop it from working? What information is needed about it? And this is fantastic information that we’re now mining as far as AI is concerned.
But this information we’ve got for 12 safety critical assets and this is the quality of information, for example. This is describing in detail what a fire door does, what risk is it mitigating? Also, how do people stop it from working? And what information do we need to be capturing during the design and construction process? What we’ve also done is we’ve had hundreds of people input to this which is a standardised way of doing fire door inspections. Or a standardised way of also doing the same with fire dampers and cavity barriers and things like that. So the principle that we’re working on here is to really share knowledge and capture it in a way that we can then use.
We’re also because of the group that we’ve got engaging with regulators. So this is something nice that the MHCLG said about us and also the HSE. And now we’ve got conversations going on at the moment with the fire service to be able to look at all of the overwhelming amount of legislation that people have got and are trying to deal with. And what we said 6-9 months ago is really coming home to roost. For example, this was before Christmas where the developer here, they couldn’t get Gateway 3 approval. Largely because of design faults, but also because they couldn’t evidence what’s been going on. The cost of this is millions.
Here’s another one that cropped up last week Lendlease are taking action against Balfour Beatty on a building that was built 25 years ago. And last week also there’s the Mace action that’s just happened where it’s £5 million about 132 fire dampers. Because they’ve been installed in a way of…which walls they’ve been installed in. So the key thing about that is I think we’ve got a business case there to really get people to start taking things seriously. Finally the elements of the promotions that we’ve been doing. We were at Digital Construction Week, so we had a team there. We also did a major event with UCL for all of the universities.
The Smart Building Conference in Birmingham and at London Build we were in 4 different panel sessions on the fire safety stage. One was with Riskbase, another with Onetrace, another with Plan Radar and another with SE Controls. So there’s a lot of knowledge that’s then being promoted through there. Finally, all of this only comes together when you actually apply it to real projects. And what we’re trying to do is identify real projects that we can take these concepts, the things that we’ve been collaborating around, to work on that.
And we’ve got a schools working group and we’re planning to have a meeting in about 2 weeks’ time as a follow up to that. We’ve also got a group looking at healthcare and generally we’ve got a group looking at digital handover which is far more generic exercise.
GORDON PRINGLE Can you just explain the claims that are coming from Lendlease and from Mace that you mentioned. Presumably they’re just following the old lines of latent defects in the sense that it’s not just a defect during construction, but it’s a latent defect over a longer period of time. I take it that’s the grounds that they’re making the claim.
GEORGE STEVENSON Yeah they’re looking at the Defective Premises Act I think, which was changed so that the…my understanding is that the building was built by a contractor that was employed by Lendlease. Then Balfour Beatty bought that contractor probably 5 years after they built the building. And now there was a problem with compartmentation and they’re holding Balfour Beatty liable. So I think the point is that what we often hear is why should we bother with the golden thread because nobody has been prosecuted. But I think we’ve got clear evidence there of at least 3 projects where that isn’t the case.
So I think that I’d like to gather more examples that that, so my request to the community is let’s try and gather as many examples of that. Because the challenge that we’ve got is that within this group we’ve got a lot of people who are already converts. The challenge we’ve got is actually persuading senior management that this is a real risk.
RICHARD FREER I’m going to share my screen and whizz you quickly through the website. At the top upcoming events, that’s just a few lines on what we’re doing and this is what we do, bringing together professionals across construction and asset management. How do we do it? We do it through our working groups that look across all sectors and they will identify specific issues that need to be covered. They will pass those through, they will set up workstreams. We have 4 meetings of these workstreams, these are task and finish. The idea being is we have something to do, let’s do it and there’s an output.
If we look at last week’s Awaab’s Law you can see our leaders there. We have meeting notes, the full video of the meeting and a 10-15 minute podcast…they’re really very good and it gives you the full gist of what happened if you missed a meeting. This section is on the Golden Thread Taskforce. The Knowledge Hub, our Knowledge Hub is you and it’s our meetings and our outputs from those meetings. There’s a tab for past events and this which goes through to the Blackbox. The Blackbox is where you can report or share, anonymously if you wish, incidents.
This would be a major incident that happened that you’re aware of, that you were a participant of or came up with a solution for to give when it happened, what happened, a photo if you’ve got it, what kind of building, short term implications. We’ve only just relaunched this so there is not a great deal on it. But if you’ve got an incident you want to share that will bring learning to the community and the wider industry put it on there. This is for asset risks, we’ve got smoke seal on door damage when using a non-closing fire door, how it was fixed.
So this is about sharing issues, whether they’re big issues or smaller issues. Then down at the bottom we’ve got resources. These are our fire critical assets here, AOVs, fire doors. This is guidance that we reviewed the middle of last year. Places to reports, digital records. So that is the website. If I just focus on one of the meetings that we had last week which is the digital handover workstream. That was looking at why digital handover information is routinely delivered but rarely used. And the next steps from that were to form sector specific working groups, schools, healthcare, residential.
Which we’ll be asking you if you wish to join any of those sector specific groups: schools, healthcare or residential. Develop practical audit and validation workflows, quantify the financial and risk value of useable digital handover, which is something we are actively working on, and produce repeatable client-led models that can be embedded contractually. So that’s the kind of thing that happens in one of our meetings. Now I’d like to hand over to Kevin Phippen who has got a good example of a specific issue that he has had. And we’ll be looking for others from you as the meeting progresses.
KEVIN PHIPPEN (shares screen) I was just trying to illustrate that I think there are lots of great things about digital information, lots of great things about BIM and the way that it works, but it can be quite confusing if you’ve not been familiar with it before. Because it becomes a bit like old McDonald had a farm, EIR, OIR, AIR, all these kind of acronyms that flow around BIM. And there’s an awful lot of data that’s involved in the handover from the end of a construction project through to the active life of a building. And I want to touch a bit on that and my experience of it.
And I think what I’m always trying to say to people is very much keep it simple. The information that’s required to maintain and look after the building once it becomes a live building is not all of the stuff that’s within the PIM, the Project Information Management files. A lot of that stuff tends to follow the Pareto example, 80% of it is there, it needs to be there, it’s information that you need. But it should be in a dead file, in an archive file. And what you need to do is separate the 20% that actually you do need that needs to go into the CAFM system from the 80% of information that’s about the construction phase. And that’s what this slide is all about.
It’s totally in line with the 19650 standard, that would encourage you again to separate things and to thin out all of the data before it goes into the CAFM system. And if you do that you’ve then got half a chance of creating a proper active CDE accommodating environment where things are being exchanged in real time. Otherwise if you don’t do that and you try and put everything, the construction files into the CAFM system you create a massive human in the loop situation where you almost need an empire of people just to keep swapping information from one part to another to ensure that it’s all up to date and compliant.
I use the example very often that a building is very similar to a car. We may not be familiar with buildings and building maintenance, but certainly most of us either own a car, drive a car or know what a car looks like. And the level of detail that’s required to maintain a car isn’t in the information that you get when you buy the car, it’s in the maintenance plan that’s held by the service station or garage where you take it for its maintenance. So you don’t need a lot of level of detail other than the car and its registration number, its VIN plate. All of the level of detail you need to maintain it is within the maintenance programme.
And it’s exactly the same with a building. You need to know from a CAFM or CMMS point of view that you have a large air handling system, you have a fire system, you have a heating system, you have cold water systems. But what you need to do to those systems and the frequency at which you need to do it should be in the maintenance programme and should not become a series of individual assets in their own right. Otherwise you end up creating massive amounts of unnecessary work and confusion and it’s very, very difficult if not impossible to keep all of that information current and up to date and to create a proper working CDE.
And here’s an example of the kind of maintenance plan we would put together for someone which covers off the main systems and explains where the LOD sits. All the stuff from COBie is relevant, but it should be divided between the active needs of the asset managing it going forward and the static or archived needs that should sit somewhere in a file. You can easily go and find it, but you don’t need to have it in the CAFM system, CMMS system because it’s not relevant for the day to day maintenance of the building.
It’s very much about the KISS thing, keep it short and simple or keep it simple, stupid. That’s a phrase that actually came from the US Airforce and was kind of in response to the COBie data from the US Army when they were developing aircrafts and maintaining aircrafts. So all of the things that I’m talking about have a long history about thinning that data down before you transfer it into the CAFM system. If anybody wants more information contact me through George and we can set up a separate meeting and we can run through it in more detail.
PIERRE VENTER Kevin, thank you, I’m very interested one of the things about assets is that we need to package the right amount of information at the right time for whatever the purpose is. And that’s one of the real beauties of digital work is in fact to be able to do exactly that. One of the things is that if we package those like we would, I use an application called Planally, but essentially what it’s doing is giving those bitesized chunks that means that the design team are populating things properly.
The handover then can take place properly and from your side on the CAFM side is that you’re getting the right amount of information. And I think sometimes we’re just overworking it. We need to actually make these little lenses that we can see and it will markedly change the way that we work.
KEVIN PHIPPEN Absolutely Pierre and just touching on that point that was brought up in the last meeting. I think it overwhelms the FM company or the service provider if you try and give them too much information. Because it doesn’t fit into their CAFM system, they don’t really need a lot of that stuff and it’s about managing the data and making it relevant before you try and hand it over. And then the handover becomes much more practical and much more useful. And the service provider wants that information, needs the information and doesn’t become overwhelmed by it. And it’s very much that process of thinning it down and making it relevant.
PIERRE VENTER Certainly 100% about that is use the right people who are actually responsible for making those decisions. Get it 100% right and then it’s available for everybody thereafter. Because they still in fact remain the owner until they’ve handed it over which is also interesting.
KEVIN PHIPPEN I think also once you get to that situation it becomes more manageable in terms of keeping everything up to date. Because one of the real problems of all of the data that’s now available is that it’s very easy to become noncompliant because you’ve got masses of data and you need to keep all of that data current. All of the O&M data, all of the historic maintenance data. And if you’re trying to put too much into the system it will fail for the wrong reasons, it won’t be updated properly and then the whole thing will become irrelevant. So it’s very important to get it right at the outset and it keeps it easier for everyone to do that than it is to try and do it after the handover.
NICHOLAS NISBET Kevin, you seem to be suggesting that COBie is using or offering data for use cases that aren’t relevant. What are the use cases that you’re rejecting?
KEVIN PHIPPEN Not so much rejecting as redirecting. I think all of the information is relevant, but from a maintenance or asset management point of view, as an example I need to know that in that building is a large air handling system. But that’s pretty much it, I just need the detail that it’s an air handling system and where it is and where I can find the various components that require maintenance. I don’t need to have all of the details of all the spigots, all the lengths of duct work, all the various vents that have been fitted to it. It’s great that I have that and put it away in a file somewhere, but I don’t need it for active asset management.
NICHOLAS NISBET That’s never been part of COBie. I think you seem to be fighting against a mythical COBie that was never proposed.
KEVIN PHIPPEN I guess it varies in all situations. Certainly my experience has been with the handover quite recently that all of that data was in the COBie files.
GEORGE STEVENSON I think the thing is Nick it’s a well-made point that people keep thinking of COBie as being the definition of the data set, whereas actually it’s a data schema to deliver whatever level of detail is agreed.
NICHOLAS NISBET It’s always been focused on maintainable equipment. I’m just trying to understand what it is that Kevin is reacting against.
KEVIN PHIPPEN In the general sense it’s simply that there’s too much information that isn’t relevant to the active asset management of the building. It might be relevant to the archive file. But I guess it’s a feature of not having an accurate employer’s asset requirement or information requirement at the start of the project within the RIBA stages. Often all of the stuff coming through is all of the detailed information that is relevant to building something but isn’t relevant to maintaining it necessarily afterwards.
DUSTIN WINGROVE I just wanted to agree with what Kevin is proposing because we issue health & safety files often, we’ll actually review information that the principal contractor will hand over to the client at the end of a project. And I think the problem stems from the contractor getting all of this information from all of the different disciplines and the contractors onsite. And what they do is they take it and just dump it into a platform of Dalux or Zutec or whatever, all of this information.
And the problem that I’ve come to realise is that if the people that are managing this information don’t know what the information is and they just put it into all of these different folders. And when I come and have a look at it I say where is the access and maintenance strategy how to I get the air unit on the roof? The information is there but it’s not actually arranged in a place where it’s actually user friendly or where a facilities management team can go where is it. In an ideal world I’d open up a model, click on it or show me it’s on the third floor. What is the safe route to get to it and what are the risks associated with it?
So that's all the information the facilities management need is to know where the item is, what are the risks associated with the item and what is the only information relating to that item? Where is at the end of a project contract to get so much information, they finish their job, they've bought the building. And they just dump all of this information on the platform and nobody takes charge of that information when they’ve handed it over. And I think as much as much as we can try and do on these sort of platforms is to arrange this and tell them this is what we need.
Until we get a key duty holder saying look, I’m responsible for this, I’ll manage this information when we hand it over, we’ll always have that problem that it’s not being managed because the client doesn’t know what half of the information is that he’s been given. He’s got all of these models that he doesn’t know how to use and stick it on an online platform and somebody says oh, I’ll look after your information. And that’s all that it is a dumping ground for all of these folders and they’re not actually putting them in the right place so that the next person comes along and uses it properly.
We have jobs all of the time where we say well, where’s the health & safety file? The building is not even 5 years old and they go I don’t know. Or they give you a dozen health & safety files from each individual contractor which defeats the point.
PIERRE VENTER Dustin, the problem is that you need to define at the beginning of the project who gets what, who looks after it, who hands it over. There’s an audit process, they hand it over and at the end of the day it gets given to the client, the developer, whoever is owning that building. And they have their mandate saying to them what they’d originally required at the very beginning of the inception, they get this stuff and that’s it. It’s a very clear process that we need to actually make sure we engage.
RICHARD FREER …I’ll hand over to George to look at further projects in digital handover.
GEORGE STEVENSON One of the challenges that comes out of what Kevin was saying there is that what often is happening, even on a brand new building…that people are sending out people to survey because they’re not getting the right asset information from the process. So findability is a really important thing and it’s a very important point that Kevin has made there, that you’ve got different stakeholders who need different information. So the CAFM system, for example, is only used by certain people within the asset management process.
Other people need lifecycle information, for example, or building safety information will be looking for different information sets. And that all needs to be delivered in some way, but they you filter out the stuff that isn’t relevant to you. And that’s why the digital side of things is important. The reason I’ve put those people up there is these are all software companies who we’re working with who are part of our community, who are willing to work collaboratively. And later on we’re going to talk about collaborative projects. So if you’re interested in collaborative projects and maybe providing some hang on to the end.
So what we’re really doing with digital information from an asset management point of view is turning record information from dumb to smart to make it machine addressable. From passive to active, from human actionable to machine actionable and basically providing the building blocks for AI. And one of the challenges that we’ve got is the way that information is typically handed over you might have a BIM model, CAD data, COBie data. A lot of the information is probably in O&Ms and 360 photos. All fantastic data, but they’re often not connected.
So the result is that the only way of checking and processing them is to do it manually. And Tim is going to show you in a moment the sort of things that the teams are doing to actually automate some of that process to remove silos, make things more connected and therefore more findable. So because we can then take the information, which is probably very good information that’s in the BIM models, but it won’t be complete because the detailed M&E information of what has been installed seldom put in. That normally comes through technical submittals and the balance of that can come through 360 photography. And that makes it a lot easier to keep up to date.
The 360 photography is an area that we’ve got a workstream working very actively in that and we’re applying it to lots of different projects, to be able to add asset intelligence to the 360 photographs. And by doing that it means that we can use 360 photography to navigate to information that is captured during the normal delivery process. We can use any camera, so it doesn’t really matter which technology you use to create the 360s.
And then there’s a whole range of fantastic applications that construction teams are using for their evidence gathering which we can use and apply from a data capture perspective to validate what’s been installed, to even do condition assessments. So basically there’s a whole range of applications that 360s can then be used to interface with.
RICHARD FREER Next is a discussion around data engineering. Let’s look at golden thread benefits…What you’ve seen already from what George has said about digital handover on projects, have we got any discussion on that? Anybody got any comments so far about anything?
GORDON PRINGLE Has anybody got a fantastic example of an exemplar digitally handed over building?
MARTIN JORGENSEN I was involved with the MEP data control side of the Elizabeth Tower project. That one was particularly well done digitally, although bear in mind that the cost involved in that amount of data management was over and above what you would expect to see in an average project. Obviously the scale of the project was enormous and every piece of masonry, every piece of equipment that was taken out of the building to restore was labelled and tracked and sent halfway round the world and people could check on its progress at any time.
And then when it’s all been put back together, obviously the amount of data that was held on each individual piece was probably to the extreme. But still it was then all put into the appropriate systems for tracking and asset management. And I believe from there, because I’m no longer involved in that project, that it is a good case of how it could be done.
RICHARD FREER It would be interesting Martin to be a fly on the wall in 5 years time to see if it’s been updated.
MARTIN JORGENSEN Very much so, that is the kind of key that I’ve seen with most projects. Because I’ve worked on everything from the extreme bargain basement cheap end of the spectrum, all the way up to huge scale public buildings. And the one consistent thing that I’ve found throughout is generally speaking when the project demands value rather than it being…effectively if cost is the primary choice as to who you pick to do the job then generally hand in hand with that you’ll find that progressively the data aspect of it is worse and worse.
JAMES FRANKLIN Just to relate to government contracts and handover. We got exemplar projects in line with open standards where we delivered massive volumes of data in accordance with the requirements. And we leveraged that data during construction effectively. But when we try to translate that into operations it’s a bit of a disappointment because a lot of that information just isn’t used. We got operators who come onboard late, if it’s a new build potentially, and they bring their own data schemas. A translation needs to occur and the chain is broken effectively and a lot of the data is effectively binned. Also the translation costs a lot of money as well.
Got FMS 002 government standard, very sensible. If you match your BIM requirement to that you can reduce that requirement by a significant amount. I’m not saying it’s the be all and end all, there are other considerations that you need to make, but that should be your base point to start from on government contracts. And you can reduce BIM, take the strain off of the supply chain, there’s a relationship between quality and quantity. And then the operator, and if you speak to the likes of Sodexo etc they’re quite happy with FMS 002, that’s a good standard for them to follow.
And then also maintain going forwards. So there is a simpler way, we’re making this really hard, really stressful, it costs a lot of money and there’s a lot of waste. And I’m not saying FMS 002 is the be all and end all and George and I have had a bit of a discussion around that, but it’s a great starting point.
TIM AIKIN (shares screen). Effectively our experience is….what we’ve learnt about constantly getting the gap in handover and what we’re trying to do is bridge that. So an example that we had, obviously if you get this gap you’re going to get these problems. They’re going to be further down the line and you’ll find them when you find them. One of the gaps that I wanted to mention, here is a COBie data set which is supposed to be a handover data set, but unfortunately it was taken from a federated model that was not updated by installation contractors.
And then at the bottom here you’ve got the installation contractors door sets and here you’ve got the door sets that are in the COBie file. The COBie file doesn’t include nay door numbers, so that makes it a bit challenging to compare it with the actual schedule. And the number of door types in the COBie is about 20 or 30 and the number in the manufacturer’s is 100. So you start to get these mismatches. The other thing is…it makes it impossible if you can’t match up by door numbers, so issue 1 there. And that’s the illustration there, there’s 7 door types that the manufacturer of the doors has actually expanded, but there’s only one type in the CObie data there.
The other thing is the manufacturer’s schedule is baked into a PDF. And if ever you’ve tried to run through a spreadsheet and a PDF file you’ll know what the problems are there. So why is it distributed like that? I don’t know. But obviously if you’ve got the right extraction tools you can extract it back out into a spreadsheet. And that is not necessarily clean and easy at times, it depends on what somebody has done to a spreadsheet. And another door manufacturer on the same project has done the same thing, embedded a spreadsheet into an O&M and of course it’s a different format. And you’ve got to extract that and sort that out into a spreadsheet as well.
And then when we came to look at the building model from the architects we found that the door numbers, where they did put the door numbers, some of them had been copied up from different levels, but in the nomenclature of the door number it actually specified what the floor level was and they’d not changed it, so you had duplicates of the door numbers. So nobody had checked through that. So these basic things to me, but obviously nobody is checking, nobody is picking it all up. So then as a company we have to do engineering gymnastics to try and restructure the information. Not something you want to do, where is the liability, and you shouldn’t have to do it.
In our spatial model which we call the pin, and then when it comes operational it becomes the A, we hold everything in a spatial database and we’re able to display it. And we’re able to locate items and then we click on the item and you get the relevant information about the instance and about its products. And about any other attribute data, maybe it also links to audits and schedules etc.
But we can also put in backdrops. So we can put in a number of backdrops and then when we look at the backdrops, here’s one which is the fire escape and they’ve very helpfully identified that those doors that are fire doors. Unfortunately they haven’t put the numbers on. So how is anybody supposed to relate that back to some of the information we don’t know. But obviously we’re able to overcome that if we’re able to multilayer information up.
Then if we get the manufacturer’s door schedule in, providing we’ve got keys that match up the door numbers, then we’re able to bring back the record of that spreadsheet. And we’ll display it in an accordion style format rather than trying to look through 50 columns on a row, we’ve just made it easier to look through.
So I looked back at this and went hang on, people are getting this massive data dumps usually as a Wetransfer. So I just wrote a programme that allows us to quickly scan either the zip file or the folder structure and give us a report. And it also gives us a navigation list so that we can then…click on the name of the file and it will just take us to the file. Here’s some examples, my favourite one is line number 4. The number of files with zero size, 35. So in the dataset, there were 35 files that didn’t have any size to them at all. Why are you providing this? Other checks are what are duplicates, whether you’ve used a certain naming convention.
And then unknown file types. Why are you providing file types that nobody knows how to open on the application? What’s the point? And then there’s a great one down there which is the spreadsheet that’s got a macro. Now I don’t know about you, but most organisations put a really big rule that nobody is to send spreadsheets to other organisations with macros in them that are a potential hole for injecting viruses into systems. So why are you providing it in a dataset? I don’t know. And the last one which is the zip files. And I call these Russian dolls. So, zip files within zip files within zip files in the dataset.
And when you actually open them up and look inside them they’ve already provided that information. So this is just to me, is it sloppy? Or is it a fact that they can’t actually check the information before they send it? The project we actually were asked to do was to check the validity of O&Ms. So we created a process where we could take the O&M and we could process it and check it. Now we already have a manual way of our teams checking things and they have an instruction set. So we took the instruction set and just simply turned it into something that the AI could read and process.
So you take the document, you have the instruction set, you send it to the AI and it then generates reports and then it gives you a summary. To process each one of those the AI takes about 15 minutes, so you just leave it on a batch run and 30-odd O&Ms done then gives you a report. There’s an awful lot of red in there, no surprise there. And this is what the physical report does, so you can send it out to people who say right, there’s why we’re passing or failing it. And when it does pass it the instructions says evidence what you’ve found. So this is where it’s put in comments and saying I’ve passed this because you’ve included X, Y, Z and there are the pages that it’s on.
The last thing I wanted to touch on is a lot of people are worried when they hear the anecdotal talking and discussion about hallucination and is it going to do this, that and the other. And there are tools already in there in the whole process. And one of them that you’ll hear about is called the RAG database which is basically where you have sets of documents that you can put in and it becomes your knowledge database, but it chunks it and turns it into vectors. And vectors are numbers and it…means it can make associations. So when you do a query you actually get candidates for then generating responses back, but it only uses that information that you’re sending it to create it.
So it’s not going out onto the internet and pulling out somebody’s…a South African building regulation or whatever, it’s using what you said. You hear about people, they set up the building regs, they set up some standards or whatever, different knowledge bases. And this really is the sort of core of things and if you start…mixing them with spatial databases and then other environmental databases you pull together a system that is basically turning the whole thing into a very smart building.
GEORGE STEVENSON What we’re offering to people, this is part of what we’re doing with the Golden Thread Taskforce which is a collaborative effort. So if you’ve got a project where you think it would be useful to run this against, to be able to do audits. Let’s say for example, for checking Regulation 38. Is all of the information there needed to satisfy building regs? If you just put ‘project’ in the chat then we can pick up and talk to you about that. Because there’s opportunities for us to apply this on live jobs, we’ve got capacity to look at that as a collaborative team.
RICHARD FREER George is now going to go through the benefits of using the golden thread.
GEORGE STEVENSON (shares screen). So what we’ve been doing over this last 18 months or so, we’ve been asking people at the various different forum meetings the current state of the art in terms of what information they’re getting. And importantly what their senior management think that’s happening. And in most cases what we’re finding is that the information managers and the subject matter experts are probably getting fairly switched on to what the golden thread is about. But as we get up the management level I think the detail just isn’t there yet, where people think that they can transfer risk whereas actually they clearly can’t.
So what we want to try and do is make people more aware of the importance and value of the information and the golden thread. Now, people throw the golden thread term around quite extensively and everybody’s got their own golden thread, in many cases. We’ve been doing this now for 5 years and really it’s important to just have some key things. It's electronic, it’s for the whole process and it also needs to be easily transferred when there is a change of ownership. I suppose at the end of the day…what we really need to do, rather than t just being a tick in the box, we need to identify what it’s actually doing for you.
And what we’re really trying to help the information managers in the community to communicate is that what it’s actually doing is translating the needs of the client into the detailed needs for design and construction and operations and recycling. And the analogy I would use is that it’s a DNA that’s working through a series of solutions, all of which work together. And the only way that projects work is if everybody is working collaboratively. If people are just looking to transfer risk we’ve got those traditional silos.
So some of the members of the Taskforce have been working on this for the last month or so, to say what information is needed? Who’s gathering it and how? Who’s benefitting from it? And importantly what are the potential cost savings? So why are we creating better quality information? What’s the business case for it? And what we’ve done is broken the process down into 9 stages that cut across whatever procurement route people are operating. And also any differentiation between disciplines. So there’s the stage of information capture, going through the process.
And what we’re really looking at is from a point of view of overall there’s planning, creation and then operate. And the common DNA there is the thread of information. So what we’ve tried to do is look at what activities that people are carrying out. So during a project there’s a series of steps that the golden thread can help with i’s tied back to activities. And the same with handover into operations. But what we’ve then tried to do so that we can start to put some tangible value against these things, we’ve then related that back to outcomes. So looking at the various different activities, what’s it actually doing? One is improving the information quality and auditability.
Reducing risk, whether it’s cost or whatever, improving supply chain. And we’ve then looked at this from a point of view of these 6 outcomes that cut across all of those activities and all of the benefits that actually come from them. So the important thing here is that the information, you’re probably already paying for it to be collected anyway. The key thing about that is being able to repurpose it and make sure that it can be presented for that value. So where we’re going with this is to say right, from a stakeholder coordination point of view there’s series of different benefits that we’ve got through having better information requirements, for example.
The same with using tools like Morta for capturing the information and classifying it properly. And as James was saying earlier, looking at FMS 002 and the other government standards for specification make a huge difference. Looking at product data, one of the benefits of having digital product data. The key thing for me is you can also make the O&M production process. And why are we validating the data? Well, if that can be automatically checked then any tools that are doing that, it means that we’ve got something that’s progressively being validated through the process.
Then we’ve got the creation of the asset, being able to bring together all of the models, the various different asset data sets, 360 photographs and then improve the process. So the reason I’m going through this is what we’re doing is looking for people that are interested in this level of detail who are willing to take some time to go through it and improve it. And also relate which software applications they’re perhaps using to see how they are then performing in this whole process.
So these are the sorts of steps that we’ve got and we’d like people to participate in this knowledge sharing so that we can actually build the value propositions for why is it that we need the maintenance information, for example. Because in many cases the construction teams aren’t aware of why it’s important to have that information about work instructions, for example. Because they assume that SFG20, for example, is a silver bullet that sorts it. SFG20 is really important, but it needs to be mapped against products and asset types so that that information can go into whatever the work or the management system is
So overall what we’re saying is we want people to be able to take this onboard and then maybe use this to cost justify why the organisation needs to invest in better information. And come up with some examples. These are some examples that we’ve come up with in terms of if it’s taking somebody maybe a couple of hours to find information that then has an impact. Lifecycle cost and risk is a significant one. Making it so that a main contractor can be more confident of the information that’s coming from their supply chain, that’s an important one.
OS the long and short of it is that what we’re looking for is input from the collaborative team so that you can map what you’re doing against this process. So if you’re interested in the benefits side of things just put ‘benefits’ in the chat.
RICHARD FREER So if anybody is interested in getting involved in the benefits, identifying them and hopefully benefiting from them themselves, put ‘benefits’ in the chat.
NICHOLAS NISBET George, isn’t there a gap at the beginning of your process of identifying what risks you want to manage for the golden thread? I mean I know we’ve been told that structural and fire are the risks we have to manage, but you can’t have information requirements unless you’ve got a purpose and the purpose is to reduce those risks.
GEORGE STEVENSON I agree and the way I would say it is that I’m pleased that you’ve raised that because as far as we’re concerned the golden thread is not limited to structure and fire. From a legal perspective, from a statutory perspective, it may be, but the philosophy of the golden thread should be related to all asset, all projects really.
NICHOLAS NISBET So yes, I mean there are obviously health & safety risks which is where we started, but there can be commercial risks or environmental risks in some projects which you may or may not want to track to the same level. Social risks like health and those kind of things. So I think the mainspring for all of this is not so much cost as risk.
GEORGE STEVENSON Cost is a risk, isn’t it? So absolutely, And I think the important thing here is to try and put some value on that process.
RICHARD FREER If anybody has got any questions for Tim also on the AI and data engineering now is the time to ask.
JAMES DOUBTFIRE Just touching on that George, because I’ve consulted on and built a few in terms of digital twins, there’s a few bits that you touch on there which are paramount. And it comes back to Nick’s point there really was well around golden thread. You need to start with your use case, so why are you doing this. And I feel your pain George, we go through this a lot, you’re trying to find a work to a commercial benefit to the client, because this is what you’re ultimately trying to sell into them, and sell sometimes sounds like a dirty word.
But you are because ultimately you’re having to build this out and map it and work your way through. In the future maybe I can run you through some of the ones that we have, but we did some for a site of ours which we built in 2019. We handed it over to the client and last year undertook twinning that environment as part of the exercise. And that was extremely problematic, even though it was a site that we thought was fairly clean. We had all of the drawings, we had all of the information, we felt like this should be something we could go in.
It was difficult even in that because even though we were on site, it’s under a managed service contract so we are doing the engineering, we are doing the maintenance, it’s the whole scheme. It’s a challenge, it’s particularly difficult and I think you’re raising some real good points that it sounds like everyone on this call is ultimately a convert anyway. So that’s great, it means that I’m speaking to people that know. But we certainly suffer when we’re talking about digital handover information twins is that we are talking a language that we all understand. It doesn’t necessarily always translate to how the operation and the maintenance of the building work.
In my past life I would put together O&M manuals and I’d get disheartened at the thought that they’d just sit on a shelf and never get touched again. But in reality if something breaks they call someone. They don’t go to the O&M manual, which they should, or in this case if it’s digitalised they should be able to go on and check and find it. But now we’re trying to bridge the gap between what we’re trying to achieve digitally and making it work in the real world.
LIAM WHEATLEY I work as part of the safety/asset management team at a housing provider. What we’ve done with the golden thread is kind of revolutionary to a point. The way that you guys are looking at it in terms of the digital twin golden thread information, we’re kind of at this point now that we’re proving why it’s so important to have that use case of golden thread management. So we’ve gone to the direction of contextualising the sites, how it would be for a customer to live and breath the site. The social value, the healthcare, the safety and overall the better asset management.
The point is we’ve saved up to £250,000 in about 6 months just through access spearheading that right information. And it was a point that Kevin made earlier about the fact that having the right type of information. And one thing that my team likes to do is to have correct information with nothing too much for people to delve into. Another point of it is DMC, that really hit us hard. And the point is that ventilation, all of that information, although that should be passed over we’ve got a lot of old, old buildings which suffer from 1960s regulations. So having those retro BIMed, having those brought in gives us a really good use case the evidence what we can do to negate DMC.
My overall point is there is absolutely for you guys use cases, there’s absolutely savings to be made. But the larger picture is it’s a springboard to something much better. Because we’ve used information at the moment to proactively look at trend data and work through what we can do with customers. But there’s so much more we can do if we AGI or AI. Everything just to give assurance to the company as a whole. If anyone wants to talk to me about that separately drop me a line.
HOWARD HALL It’s just a quick reminder that when we’re looking at risk, we’ve obviously on any project that’s like a construction project, refurb, whether it’s through the FM activity, should always have a risk register anyway. And it’s quite surprising how often risk is only annotated to the design or the construction and not necessarily the effects of a poorly operated or poor digital or poor design. And therefore we’re trying to now, we’re obviously including the risk of golden thread or lack of information or the integrity of the digital information flow within the DRR, the Design Risk Register.
And it’s just another prompt to be able to bring it to the attention of the client and the team. So rather than actually having to make a separate business case let’s try and use the exiting tools. The other thing that occurred to me is the conversation about business case is really valid because we’ve always got to justify any increase in service fee or cost. And if anyone has got any data which is actual data of where they’ve enabled a real time saving. Anonymise the client and the site, just give a detail of what sort of site it is. If we can start collating data which is anonymised it will then further validate where these percentages and were these assumptions are coming from.
But we’re all under the NDAs and the confidentiality conflicts, if we just anonymise it then we should be able to build it up. Because otherwise it’s still going to be another data set that we’re that we’re going to have to justify.
TIM AIKIN I just wanted to expand on James’ point which he said if something goes wrong they ring somebody up instead of going to the O&M. OK, so when they ring somebody up what we want them to do is go to the AI system. I started developing this for my own house for an AI system and if you go on LinkedIn you’ll find a video of it where I just ask a question about where are all of the doors in the house or what is the warranty on the fridge or some appliance. And it told me exactly right Ok, where the gaps are. I know that the other thing is that we never actually test data to see whether it’s fit for operations because nobody knows how to do it before you hand it over. So you’re really left with the experience of the FM people sort of coming in.
But now we have the opportunity to actually gather that knowledge together and start to build out these ways of testing the information, whether it’s fit for purpose, as it’s being created and passed through the pipeline. Changing the concept of the O&M. We did the O&M review and everything, but it showed me that what the O&M is is really a sort of consolidated executive summary of what is being provided and why it’s being provided, but in fact it’s very vague. Whereas in fact what we should be saying is we need these data sets and their products and then we can test it. Because more of it is data, it’s easy to test.
Yes, we need all of the narratives and things that explain to people how to do something or what to do in this particular case. But at the end of the day, it’s like science fiction, it’s almost like we’re living on the edge of it now and we can start to see it happening. And I think it’s going to happen quite rapidly, but ultimately the information that you put together isn’t wasted because we can show it’s actually been tested. Tested with scenarios, tested with what ifs. And the thing about the digital twin, what’s going to happen in 5 years’ time to my building, give me if this situation occurs so that we can prepare for the worst and hope for the best.
I’ve just seen that COBie is actually in the FMS 002 under Section 7 data systems. 7.2, the asset register should be as a minimum codified in line with the standards being transferred, and it mentions COBie. So there you go.
GEORGE STEVENSON What we’re really looking for, as Tim is saying there, if you’re doing things in either the data engineering or the AI front, we’d like to include you in the groups that are building those types of relationships. Because there’s lots of people using AI now, they’re using ChatGPT, Copilot etc. If we can bring together the people that are doing that and work as part of the Golden Thread Taskforce then we’ve got something that then is more joined up. (shares screen).
We’ve got a series of projects that we’ve been operating over the last few months and we’ve just had a couple more that have been added. One of them, James, is with one of your projects. Essentially what we’re doing is we’re trying to pull together people that have got tested solutions. So for example James just now, he can join. And these will be people that have either got services or software and the important thing about this is that we can encourage people or we can get them working together before they’re used on actual projects.
Therefore we’ve got trusted relationships with things that already work and we can bring those together with the type of standardisation, classified libraries and standardisation, so that we can then bring the solutions together in a way that then can deliver pragmatic and certified processes. This is one that the Taskforce is doing at the moment where we’ve got a combination of Equans, Multivista and the NHS at Sandwell. So we’ve basically taken the CAD drawings, there wasn’t a BIM model on this (at least we’ve not been provided it).
And then we’ve got a spatial model, this has just happened two weeks ago, that Multivista are now going to take 360 photos. And then those will then be used to verify the asset information that’s come through the delivery process. We’ve got a major project on at the moment with Ridge and Riskflag which is Tower Hamlets. And Watney Market which we’ve been working on for the last few months, this is now going to be scanned and we want input from people who are involved in asset capture for us to get involved in this as well.
This is something where you can track what we’ve been doing. So if you’re interested in Watney Market, there’s been some delays on it but we’re now kicking off on that. And the principle here is that we’ve got information that needs to be coordinated. So the golden thread, if you’ve got a golden thread project we need you to be able to demonstrate that it’s not just individual systems that are there, it’s systems of systems. So that if, for example, a fire breaks out in a kitchen there we want to make sure that the smoke doesn't get into that stairwell,
And no matter how good the door system is, for example, unless the penetration seals are doing properly we’re not going to get anywhere. So that’s the sort of process that we’re operating where we’re treating the asset information in a way that then can be used by any software application that’s maybe doing those inspections. And here’s a live example from a real project where the information from the construction team from the O&Ms has been reverse engineered, using the technology that Tim was showing there.
But the important thing here isn’t which software is being used, it’s actually that what we’re doing federating information that is say the fire zone information and also the information about the safety critical assets that are maintaining those compartments. So we’ve got the compartments there but also the information about where the doors are and what the specification of the doors are. So this information again is coming from manufacturers. So these are the sorts of examples of golden thread information where we can also tie it back to the fire strategy.
I was talking with somebody on Friday where by producing an updated fire strategy plan they’d been able to remove the need for hundreds of fire dampers because they’re no longer in compartments. Because the new fire strategy has moved from the fire compartment being a room to being a much bigger area. So that’s why the fire strategy plans are really quite important. So that’s the sort of process that we think we can bring if we’ve got lots of different federated information. So if you’re interested in any projects, either joining the projects that we’re already working on, but more importantly bringing a project to the table, then put ‘projects’ into the chat.
RICHARD FREER That project includes if you’ve got an O&M that you want the Taskforce to analyse and validate, put ‘project’ in the chat. And if you’re interested in our AI group put ‘AI’ in the chat. We’’ll send out invites for a quick one-to-one chat with one of our steering group. These will be 15 minutes, a quick chat to see what you want to get out of what we’re doing and how we can help you.
GEORGE STEVENSON I forgot to mention this, this is the result of our LLM that we’re using on the datasets at the moment. This is just asked a question on plain language looking a 7,000 documents, does it contain the information that’s needed to satisfy Regulation 38? And you’ll see here, it identifies what is there, which is very helpful, but importantly what isn’t there. And this takes 3 or 4 minutes to do. And we’ve done this for different data sets, for an electrical safety perspective, for example, exactly the same as well.
So because of the platform we can now do this very easily. So if you do have an O&M that you need to check or if you’ve got something coming up to a Gateway then it’s the sort of thing that can be done by the Taskforce. Obviously there will be a fee for that because different people have got to be paid, but it’s a fraction of the cost that it would otherwise be and it’s quick.
RICHARD FREER Our next monthly meeting is the 11th, same time, see you then.
CHAT
Neville Tomblin (External)…. 75% of GW2 being rejected is unacceptable. The problem is that different BSR Leads and MDTs take differing views on what information should be provided. TH The BSR needs to have a more co-ordinated approach to this to avoid this and to provide some sort of sense checking.Building a GW2 application is not rocket science!!
Gordon Pringle (External)… The car has alarms and fault systems that tell you when something is failing. A Building or building system can have similar.
Graham, Claire (External)… System-Parent-Child-Component specifications
Dan O'Gorman (External)… The BIM CAFM CDE Safety File and BMS EMS solutions and PPM Schedules (SFG20) feed the "Digital Building Logbook" (DBL) as the open interoperable host of the PERTINET information relevant to those charged with the operation and manintenace of the building
Chris Jeffers (External)… In my experience, FM companies and service providers know what they need and what they want for CAFM systems etc. but this is not mandated - or clearly set out - in the construction contract so remains an issue of uncertainty through the RIBA stages, and we end up with a commercial based "tussle" where the client will not want the project to incur any further costs, but more work is needed to define the handover information. OIRs, EIRs, and AIRs all essential, but they have to work for the end user organisation(s).
Victoria Finn (External)… we seem to get everything but relevant information against assets and have produced a matrix to be included in tenders to show the relevant information required but it seems to be too difficult to get hold of it
Andrew Simpson (Unverified)…. Format is hardly ever aligned at the initial stages and this is often completed at the latter stages.
Chris Jeffers (External)…. FMS002 as starting point - exactly right
James Doubtfire (External)… Language and frameworks is a great place to start.*common language that is
Liam Wheatley (External)… if a common data schema hasnt been agreed, language is always a risk to data handover, it could be aligned with AI workstreams to identify different descriptors.
Ron Hedger (External)… Tim: I find this all the time when trying to develop a fire strategy. What I need is a very small sub-set of the full data, but if I have to try to find it all and verify its correctness, this can take me a huge amount of extra time.
Phil Hannabuss (External)… What "AI" is being used for this, and wondering how is the data being secured?
Alan Brown (External)…. Hello Tim, this is Alan Brown from Fire DNA, with regards to door and ironmongery schedules and the import / export into our software , we have undertaken a body of work , shall we chat about our joint experiences
Matthew Crew (External)…. Apologies all, but I need to drop off now. Fascinating discussion. Would love to join some of your working groups. Thanks very much.
bernard hornung (External)…. The visual only inspection has long been a ritual in checking buildings. A new wave of technology can measure and reveal invisible risks, and a building walk through might soon be powered less by human eyes and more by a combination of sensors. I am interested in joining the AI group. Thank you.
Chun Qing Li (External)… George Stevenson I'd like to participate on behalf of KREODx
Nicholas Nisbet (Unverified),…. RK_50_40_26 Failure to deliver data and information
RK_50_40_28 Failure to exploit data and information
RK_50_40_35 Failure to produce required data and information
Gordon Pringle (Unverified)… Howard I may be able to assist with operation system device data so maybe a chat sometime.
Cameron, Ross (External)…. Whats the software you are using here to show these plans please?
Felipe Manzatucci (External)… We'd be keen to get involved in AI, Golden Thread, and Projects.
STRUCTURED ACADEMIC SUMMARY
1. Forum Context and Purpose
The BIM in Asset Management community reviewed five years of BIM4Housing learning and two years of BIM in Asset Management work, with a focus on how structured, standardised information can support safer building outcomes, reduce operational risk, and enable “Golden Thread” compliance. The forum emphasised practical application across real projects and the translation of legislative intent into consistent, repeatable industry practice.
2. Golden Thread: From Ambiguous Legislation to Standardised Practice
A core theme was that legislation often lacks explicit operational requirements, creating inconsistent interpretations by experts and confusion across supply chains. The Taskforce approach is to resolve this by collaboratively building standardised information requirements, data templates, classification structures, and reusable libraries that translate regulatory expectations into practical deliverables.
3. Safety-Critical Asset Knowledge Libraries and AI-Readiness
The forum highlighted structured outputs for safety-critical assets (e.g., fire doors, fire dampers, cavity barriers), including:
- what risk an asset mitigates,
- how it fails (including human behaviours that disable performance), and
- what information must be captured through design, construction, and inspection.
These datasets are increasingly valuable as training and validation inputs for AI-driven checking and evidence workflows.
4. Emerging Evidence of Legal and Commercial Exposure
Real-world examples were used to strengthen the business case for Golden Thread adoption, including disputes and claims linked to compartmentation and fire safety defects (including fire dampers and installation/placement issues). The forum positioned these cases as tangible evidence that liability and enforcement risk is becoming more visible, and that “risk transfer” assumptions are increasingly unreliable.
5. Digital Handover: “Data Dump” Failure and the Need for Role-Based Information Sets
The discussion identified a recurring industry pattern: large volumes of disconnected information are handed over at completion (models, COBie, O&Ms, PDFs, photos), but not structured for operational use. Participants argued that successful handover requires:
- clear information requirements from the outset (OIR/EIR/AIR),
- separation of “active operational information” from “archived project information”, and
- responsibility for governance at handover and beyond.
6. COBie: Schema vs Misuse in Practice
A technical debate explored how COBie is often blamed for “too much data”, when in reality it is a schema that can deliver the agreed level of detail. The practical issue is not COBie itself, but poorly defined requirements, inconsistent modelling discipline, and failure to align data to operational use cases.
7. O&M and H&S Files: Findability, Validation, and “Duty Holder Ownership”
The forum reinforced that information may exist, but still be unusable if it is:
- stored in unstructured folder hierarchies,
- duplicated across contractors,
- missing critical operational strategies (e.g., access and maintenance strategy), or
- not validated and indexed for retrieval.
Participants stressed the need for an accountable duty holder to ensure information remains curated and operationally accessible.
8. Data Engineering and AI: Automating Validation and Reducing Manual “Gymnastics”
Practical examples were provided showing how AI and data engineering can be used to:
- compare COBie outputs against manufacturer schedules,
- detect mismatches (e.g., missing door numbers, inconsistent types),
- extract data embedded in PDFs,
- scan data dumps for quality issues (zero-size files, duplicates, unknown file types, macro risks, nested zip files), and
- generate automated pass/fail reports for O&M validation.
9. RAG (Retrieval Augmented Generation) and Secured “Closed Knowledge” AI
The forum discussed the role of RAG-style approaches, where AI answers are generated only from approved internal datasets rather than the open internet. This reduces hallucination risk and supports evidence-based assurance outputs for compliance use cases (e.g., Regulation 38 readiness checks).
10. Use Case–Led Value: Risk Reduction, Cost Avoidance, and Operational Savings
Participants repeatedly returned to the need for clear use cases to justify investment. A housing provider reported savings of up to £250,000 in six months through improved access and targeted information availability. The forum also called for anonymised case data to build an evidence base for board-level business cases and cost/risk justification.
KEY INSIGHTS
1. Golden Thread success depends on standardised, reusable information structures, not new documents.
2. Ambiguity in regulation drives inconsistent expert guidance, which confuses supply chains and undermines compliance.
3. Operational handover fails when information is delivered as a “dump” rather than a curated dataset aligned to user roles.
4. Findability is a primary safety and cost driver: information can exist but still be unusable if it cannot be located and trusted.
5. COBie is frequently misunderstood; failure typically comes from weak AIR/EIR definition and poor governance, not the schema itself.
6. AI-enabled assurance is now practical (minutes per dataset), enabling rapid checks for Regulation 38 and other compliance needs.
7. Data quality issues are often basic but high impact (missing IDs, duplicates, PDFs-as-data, inconsistent naming, macros).
8. The business case is increasingly defensible using real claims and disputes linked to compartmentation and fire safety defects.
9. Risk is the mainspring: cost justification is strongest when framed as prevention of safety, legal, commercial, and operational risk.
10. Anonymised evidence is essential to scale adoption and persuade senior leadership beyond the “converted” community.
KEY RISKS
1. Regulatory non-compliance risk due to incomplete, inconsistent, or unverifiable information handover.
2. Legal liability exposure from latent defects claims (e.g., compartmentation, fire dampers, door performance).
3. Operational inefficiency and re-survey cost caused by missing or unusable asset information.
4. Inability to evidence Gateway compliance where documentation cannot demonstrate design intent or as-built reality.
5. Safety-critical asset failure due to missing inspection data, unclear requirements, or poor identification and traceability.
6. Cyber and IT security risk where macro-enabled spreadsheets and unknown file types are included in handover datasets.
7. Vendor and tool fragmentation leading to disconnected models, O&Ms, photos, and schedules that require manual reconciliation.
8. Data decay risk: even exemplary handovers may degrade rapidly if ownership and update responsibility are unclear.
9. Senior management blind spots where leadership assumes risk can be transferred contractually rather than managed operationally.
10. Failure to align requirements to use cases, resulting in wasteful data production and low adoption by operators.
ACTION POINTS
Immediate (0–4 weeks)
- Define and publish minimum operational information sets (the “20% active set”) distinct from archived project data.
- Create a simple handover validation checklist covering IDs, naming, file types, duplicates, and completeness.
- Recruit members into the AI group / projects group to apply validation workflows to live datasets.
- Begin collecting anonymised case studies of cost avoidance, rework reduction, and compliance assurance wins.
Medium-Term (1–6 months)
- Standardise Regulation 38 evidence expectations into a reusable library and audit workflow.
- Develop structured “asset-risk-mitigation” templates for additional safety-critical assets beyond the initial 12.
- Formalise RAG-based assurance outputs so AI responses are bounded to approved datasets.
- Align digital handover requirements to government standards (e.g., FMS 002) and map them to CAFM/CMMS needs.
Long-Term (6–18 months)
- Build an integrated Digital Building Logbook model linking BIM, CAFM, BMS/EMS, safety files, and PPM schedules.
- Establish repeatable client-led contractual models that mandate usable handover and post-handover governance.
- Scale “systems-of-systems” Golden Thread application across real portfolios (schools, healthcare, residential).