BIM IN ASSET MANAGEMENT MONTHLY FORUM 29-04-2026
GEORGE STEVENSON We’ve got quite a big cohort that we started 6 years ago with BIM4Housing. And then about 2 years ago we created BIM In Asset Management, I think this is almost our birthday.
RICHARD FREER It is, it’s been actually 2 years and our first meeting that we had for BIM In Asset Management we had 24 people there and now we usually get certainly over 200. We had about 400 people on our membership database 2 years ago and we’ve now got nearly 10,000, so not bad in 2 years.
GEORGE STEVENSON One of the things that we’re really looking at is rather than BIM being about 3D modelling it’s about being able to look across the different working groups that we’ve got to look at what information makes the job easier for the next person in the chain. We’ve got a whole range of different tools that we’ve built over the last 5 years and what we’re also constantly doing is looking at what information is needed for different purposes.
And the objective here is to bring together people. These people, for example, are all manufacturers, these are all designers. And just in this community we’ve got a really good spread of people who are sharing their knowledge. I’m going to show some of the work that we’ve been doing with the Construction Leadership Council because one of the things that they’ve asked us help define what as-built means. So we’ve had a series of workshops around that over the last few weeks which I can share with you.
I’m going to stop there for the moment because part of the process of saying what information is needed is also to see what’s missing. And we’ve identified that information that’s missing is a risk and it’s a risk both from an operational point of view, but also from a statutory point of view. And therefore not knowing certain information at key points, that ought to be seen as a defect.
So we’ve been working with the team to identify how those things could be identified as digital defects and therefore be covered under normal contractual requirements to deliver the information that’s missing and maybe even get that paid for, but also to look at that from an insurance perspective. Mike, could you explain what we’re doing with the insurance side of things?
MICHAEL CHAMBERS I am head of construction real estate specialist risks for an independent broker called Vista, I’ve been in the construction real estate game for about 20 years. And my role is as a broker and risk advisor, I work with everybody in the supply chain, either on specific projects or for their own programmes, to identify what is going to trip them up.
And that can be anything from aligning some insured existing structures, project plan and building safety regulations…is identifying the data pertinent to that in the stakeholder and supply chain for delivery of certification, but also the insurable risk appetite for whether or not a defect in that process can be constituted as a material exposure. So I put it down to three points, is early engagement is critical. So if you have a risk partner and they know what your pipeline or project looks like, involving them as early as possible will help you to navigate and consider the risk landscape.
And it’s important that insurance isn't an afterthought. I know when you get to sort of PCSA stage, insurance credentials or clauses are operative, but I think having a risk partner to help that process is going to be invaluable. That’s been my experience and what I see is the last minute thoughts of insurance and alignment for is very hard. The second tranche is an over and above approach to data and that’s something that this group do very, very well, is how do we translate that process, that data capture into repositories that speak insurers language.
Super important is my role to present risks in the market and oversee a line with a suitable solution. And the third tranche is that route to market, how do we have insurer’s support at that early juncture? And what do they need to go above and beyond the basic requirements? A lot of risks are presented as numbers on a piece of paper, where’s your postcode, how long is it going on for, hat’s the GDV or ECV sums. And they sort of paint by numbers. What you can't do is have an all singing, all dancing insurance programme for the very beginning, you have to build like a foundation, superstructure, roof, everything.
So part of my role as that early engaged risk partner is to capture that information, present it to the market in a way that they understand, but where I explain about that route, you can’t just go to everybody and hope that somebody will come back with something meaningful. It’s working with a select panel of partner insurers, give them information that goes above and beyond those numbers on a page, that basic sums insured detail. And any broker can do that, but a real true intermediary and advisor will be the one to substantiate and add the narrative that means the risk is considered that much better.
The produce offering is that much more robust in terms of wording and responsiveness and ultimately you don’t pay through the nose for it. Because insurers will either exclude put conditions, they’ll put overinflated premiums when they don’t understand the information. And thanks to this group the information you put into the as-built plan and building safety regulation, that data I can definitely snatch bits that will go towards things like reinstatement, things like trace and access. These exposures can be quite basic and off the shelf, but you can actually fine tune quite a lot of insurance programme. This is contrary to a lot of people’s belief because insurance is a necessary evil.
So the long and short of it, what myself and George are doing is probably create a white paper and short webinar to say well, here is a project and I’m thinking along the lines of doing an example build for an insurance programme alignment policy creation and policy responsiveness. And probably a separate tranche would be to do a loss scenario. So we’d look at a build and either there has been a loss onsite or the remediation required, how would insurers respond to it? And that’s something we’ll do either led by myself and the team at Vista or with a partner insurer. It’s something that’s an ongoing process and we’d love the group to feed into that idea.
GEORGE STEVENSON One of the interesting things, Mike and I have been talking about this probably for the last 2 or 3 months now and this point about the digital defects comes in because if we’ve got projects where the information hasn’t been handed over in a complete data set. There’s all sorts of reasons why that may have happened and people can blame the main contractor, but actually the information is coming from specialist subcontractors. So rather than looking at blaming people what we’re trying to do is look at how we can help them get it right in the first place. And therefore the type of audit process that I’m going to talk about later is important.
But from a risk perspective I’m told, I don’t know how many people are housing associations and councils on the schemes at the moment where there coming up maybe to Gateway 3. If they have got in place, because everybody has gone ahead and produced their risk management plans, they’ve had to do that because of the regulations on the residential side of things anyway. But the challenge is that you may have written into your procedures activities that you’re following to mitigate risks.
And I’m led to believe that people are struggling with that at Gateway 3 because at Gateway 3 if you’ve written down that you’re doing certain things and you’re not doing them then that’s an issue. So that’s where the golden thread needs to be articulated in a way that’s joined up and therefore we can then identify what the risks are. And then if the Regulator says actually you don’t have the evidence that you expected to have, what are you doing about that, you can have in place a plan to actually work through that. Is that right, Michael?
MICHAEL CHAMBERS Absolutely, so one of the pieces that I’m writing at the moment for the Building Safety Act is around split PI exposures between design and principal design roles. So accountability is the big thing under building safety regulation, we all know that, but it’s where you are responsible for the golden thread of information where the blame game occurs because professional indemnity looks in to the past, the class of insurers called claims made. So what they’ll do is say well, who was involved in this? Who are the accountable parties for that transfer of information, the collation of information?
And then they'll look to have an insurance response. So depending on how the insurance programme is created, who is allocated that responsibility, the first port of call will be the principal contractor. And they say, well, actually, I involve these parties to ratify technical decisions and or design, then they get rolled into it. There's an awful lot of respondents in arbitration under contract. So again it’s trying to absolve the supply chain. The reason I’m writing the article about split PI exposure is to share the risk, share the burden of that culpability to reduce premium in the longer term.
It’s something I devised when I was dealing with the building safety pre-extension. But it’s very simple to align the blame to somebody and arbitration costs an awful lot of money. Mediation is the cheaper of the two, but if you’re able to completely mitigate that from being an issue through that sensible and pragmatic approach to data and have the insurers understand that culpability and throughput of responsibility you’re going to win. And then the long game will be not off of the shelf premiums, not products with significant conditions or exclusions or PI that has write outs and carve outs to specific exposures.
It’s an ongoing process, but there are insurers certainly onboard because the more information you get the better the result will be. That again goes to my first point regarding engagement. If anybody has any questions I’ll put my email in the chat. If there are any projects that you’re working on I love the ‘what if’ questions: What about this? How would you approach that? I am happy to answer anything.
GEORGE STEVENSON I think the other thing is that Mike has offered to run a dedicated workshop on this on the 18th of May. So if this is a topic that interests you then if you put ‘insurance’ in the chat and we’ll get you an invite.
MICHAEL CHAMBERS And then George and I will definitely put together an agenda. Insurance is super boring but it’s relevant to everybody, wherever you are in the supply chain there’s a portion of risk that comes up in every conversation. All of the work that I did in the public sector, loads of assumptions that get made and then contracts are drafter with LAD provisions done, EOTs in there. And then you have to define what’s the employer action, what’s the client action and all of these things. So it’s about having me involved early, long story short.
RICHARD FREER Has anybody got any comments on this?
PHIL THOMPSON I put a comment in the chat…it’s really for people in the insurance industry, but how do insurers propose to trace products by the golden thread that have been delivered to a project as a product and not as an asset?
GEORGE STEVENSON I think that will be a procedural thing that the construction team will have to demonstrate that they’ve done and then the client has got to demonstrate it. And it will be up to…I guess they can then test as to whether they think that’s adequate from that point of view.
PHIL THOMPSON Have you seen the piece of work that Bolo has been doing on evaluating project risk for insurers? It may well be that it’s worth reaching out to him because he’s actually put an update in the chat from the previous GS1 working group with a link to the work that he’s doing in the working group on the framework on evaluating insurance risk, which I suspect is very closely linked to what you’re doing here.
RICHARD FREER George, do you want to get started on your digital handover piece?
GEORGE STEVENSON In the last month we’ve done a lot of interesting stuff with a couple of the different partners and one of them I just wanted to show an example of. This is coming out of the healthcare sector. The Midland Metropolitan Hospital which is a new hospital, they’ve built a learning centre and they’re going through the process of auditing that at the moment from a point of view of handover. So what we’ve done, they asked us to put together a spacial model and they’ve brought in Multivista to do 360 photography.
So the first thing that we did was create the spacial model, this was created from the drawings from the O&M information. And we’ve then got the data there to say what each of those spaces is. So each of these becomes a data container which can then identify what the functional use of that space is so that we can overlay the fire strategy against that as well. And we’re now starting to bring in other information. But the Multivista solution, we’ve got the 360 photos which then are identifying what the condition was at handover, so that’s a digital record.
And what we’re currently doing, they’ve got about 20,000 assets in this building that Mott McDonald, another one of our partners, has provided, which we’re bringing in and adding standardised taxonomy to. So that we’ve got that information to be able to then create the asset information model so it can all be tied together. Now the interesting thing with this is that the fire stopping information was delivered as a really good digital record, but it was a 4,500 page document which is just completely overwhelming for anyone to actually use.
So from a contractual perspective, from a compliance perspective, it did the job, but how you would then be able to consume that information and use it is a bit of a challenge. This is a fairly typical fire stopping report, so you’ve got basically a drawing of the architectural plan and then they’ve put a little pin down on that image and then the record of what’s installed, Really good information, who the manufacturer is, what the product is, a bar code against it. But there’s a single pin for every single fire stopping asset, so that’s why it’s 4,500 pages.
And we’ve done a similar project for Skanska on a project and in this case this was the fire stopping report. This was from another application called Boris and again it’s the same methodology where the record information is very meticulous, but the only way of knowing where that item is is by looking at an image on a drawing. So what we’ve done is we’ve taken the locations of each of those pins and we’ve scaled the drawing up to be one to one, so effectively we can then identify where every single pin was located.
And we’ve then read that data into a database, so we’ve actually got both the product and also it’s location. And then that’s now presented back as a single spacial model. So that then becomes something that’s interactive and it’s not separate items of fire stopping. So we’ve got fire stopping, we’ve got fire doors, we’ve got compartment walls all in the same environment. So it means that you’ve got something then that you can manage going forward. So it’s that type of coordination, this is actually reading back to the fire stopping document and the information that’s in the document is then in the database.
And then you can also overlay other information such as the fire zones. In this case we’ve got the fire zones that are being read in there as well. So it means that we’ve got a coordination of asset information that would otherwise be all in separate files. So here for example, we’ve got the fire zone plans actually as part of the asset information model. And we’ve also got the documents that then go to make that up as well. This is overlaying the compartmentation drawings and overlaying some of the other risks. So I think that type of hazard management is a way of de-risking the process which can then inform the reduction in premiums.
I’d now like to introduce something that we’re learning from other industries. Ian Upton who is a specialist in asset management, largely not so much on buildings, he comes from an asset management and oil and gas background.
IAN UPTON I’m from a nuclear, chemical, Ministry of Defence background, estates defence management, DIO (Defence Infrastructure Organisation), industrial and steel manufacturing Tata Steel. So having been a chief engineer and process safety manager, that’s kind of my background and I came into the estates and facilities management through a career in combined chief engineer, Clerk of Works, that type of role facilities management. And later HAZOPs and things like that in industry.
So what I was interested in, until a few weeks ago I hadn’t heard of Gateway 3, although I’d heard of the Building Safety Act and the work that Judith Hackitt was doing on reviewing that as part of the Grenfell investigation. So it really interested me in terms of HAZOP kind studies and I’ve put a few slides together to go through how I feel that those link together into asset management, how they’re linked through to the ISO standards that are available, such as 19650, 8000 for data quality and 55000 for asset management. So I’ve detailed that in a bit of a flow chart which I’ll present now.
So I put together a few slides on Gateway 3 as I understood it and how that feeds through, going on from the previous presenter’s comments about insurance and meeting the evidence requirements of Gateway 3 and what’s within the Building Safety Act. So that’s the three pillars that are brought up in terms of what I believe is needed or required to meet those requirements and then the process flow at the bottom. So you start off with your data sources, you do your validation and you put the information in a common data environment. You collect the information for your evidence pack and then it goes to the Building Safety Regulator.
So that’s how I understand it…my background is asset management and looking at the assets and the equipment and how to manage estates and infrastructure. But as this goes from my background in HAZOPs I realised that when I looked at the RIBA 7 stage process it aligned really well with what HAZOP was. So I was like we already collect this information, or I would think a lot of organisations already collect the information for HAZOP, so there’s not really a big step up to prove that you’ve done the work if you’re going to do it.
So if I go to the second slide and look at what standards feed in to that and regulations, obviously you’ve got the Building Safety Act, Regulation 38 and BIM standard and then the ISO data standard. There’s also GOVS004 and FMS002 which is the data standard for government buildings. So these are baselines as it were for me that I’ve looked at in regards to what are the standards that we need to meet to make this compliant. So those are the kind of standards that feed into those pillars, but going back to the pillars as they were it’s fed in from the information that you collect in the very beginning with the data.
Now we know it’s been an issue with handover documentation, design drawings and things like that missing. So obviously this whole process is meant to make people responsible so they collect that information from the very beginning. They do the due process within building safety case evidence and then present it as as-built information to hopefully comply with what’s required. So if we take the HAZOP as a whole process and align that with the 7 stage process in RIBA, it complies really well. It aligns it really well as well because you’re already collecting that information.
You should have already identified the hazards that you’ve got. Now maybe you have to do a bit more work on that to prove that, for example fire escape routes, compartmentation and smoke dissipation, things like that. So you should have all of that information. Maybe you’ve not, but if you haven’t then it’s something that you maybe need to consider. But also with identifying what hazards may be around the building, chalk points and things like that for example. So I thought really this is very aligned to what a HAZOP does.
It defines what the project is, it IDs what hazards may be, then it does the risk assessment about assessing how badly those hazards may be. But bearing in mind you may only have those hazards without any controls in place, that’s your first risk assessment really, what would it look like without any controls in place whatsoever? If we just give everybody free rein and let them do what they want, what would that look like? And then what are the mitigations to put in place that help mitigate that from happening or prevent that from happening? What are the barriers/
So that’s your risk assessment stage 3 HAZOP. And then you have extra conditions that you might put in place. Traditionally this would be a fire officers job, he would go in and assess a building and say well, I’m not really happy about those curtains over there, are they tested to BS5867? There in a public building, have you got your test certificate for those curtains? And then you would have to produce that. So the risk assessment is really there for those checks and balances, have we got the measures in place that say this is compliant? And that can be done at development stage.
GEORGE STEVENSON The point about this is that what we want to try and do is draw on Ian’s experience in this area and he’s offered to run a workshop for us on the 13th of May. So if anybody is interested in this procedure of aligning the risks against this methodology of working then if you put ‘risks’ in the chat then we can give you an invite to that meeting. The thing I think is particularly exciting about this is that most people, certainly in the residential sector, are probably using the bowtie methodology for creating the safety cases.
Now on those, if we can look at the digital defect as being the risk then we can align it to, as Ian is saying, the whole process. And therefore what we’re doing is we’re aligning things in a way that then becomes more joined up. And I think that’s going to be really important when we’re talking to regulators because it’s all very well having a standard operating procedure for dealing with a risk, but if you then don’t follow it that’s probably worse than not having a procedure at all. So I think that’s an important part of that.
I’m going to share…it’s part of the process that we’ve been doing with validation, but what we want to do is make sure that the information that has been handed over as the evidence, from an asset management point of view (remember that we’re talking about BIM In Asset Management), we’ve built a three stage process to operate and we’re actually running this on live projects now. We’ve been give several projects to test since last time and if you’ve got any projects that you’d like us to test for you then put ‘validation’ down in the chat and we can come back to you with that.
The steps are first of all if there is BIM data check that the information that’s in the BIM models and the exported information of the asset data is actually compliant with the requirements. So anything that’s been modelled we can then capture as asset information that we can then reuse. But probably only half of the asset information is ever put in the models because it’s coming from specialist subcontractors. And the way that they deliver their information is typically in the form of documents which then may or not be read a checked.
So what we’ve been doing over the last 18 months is building the processes and the technology to actually check that the files that were promised in the registers and schedules have actually been provided. So that’s the first step in the validation process of the health & safety file and the O&Ms. The second step is to check that the content of the work packages satisfies what’s expected. Now that’s all configurable, you can determine the level of requirement, the level of detail, the level of accuracy that you want, but then it will then check against that.
And the third area is reporting each work package in terms of what’s missing and how to fix it. And then the third part of that process, so we’ve got asset information from the models, we’ve got asset information from the O&Ms, but now we can then use the 360 verification that can be used to verify what assets have actually been installed, because you’ve got 360 photos of the real thing. And just to illustrate the way that we’re doing that is that although the BIM models often will go to the construction teams they invariably don’t get updated with what the installation contractors have put in.
So the BIM models tend to be design models that the FMs get and the rest of the information they’ve got to get from O&Ms. So what we’re doing is analysing the O&M files, so in this case we’ve got 10,000 O&M documents, and the process, the system will automatically go through and check whether the files are valid or not. In this case on row 4 you’ve got 35 files that have got zero size which means they’re empty. We used this on a scheme a couple of weeks ago and the fire strategy plans, they were absolutely blank. There was a placeholder for fire strategy plans but of this 10-storey building there was only one that was actually displaying anything because nobody had checked.
So this is helping that overcome that manual checking. And this is then going through manually checking, with experts checking the content of each work package. In this case there were 54 work packages and we’ve used automation subsequently, based on the knowledge of the experts that have done this work, to actually say well, does it have operating procedures? Is the warranty information there? Was the technical information there? And that then produces a report for each work package saying what’s passed and what’s failed.
Now my colleague Tim is going to just explain briefly what we’re doing with the other workstream session which is tomorrow on AI and data engineering. What we’re doing is we’re taking the knowledge that was provided by the experts and that’s been turned into agents which can then be run against that dataset when it comes back for checking. Because what we’re doing is we’re providing a report back to each work package to say what has passed and what has failed that particular aspect. And then we can then also use that information against the models.
So here it’s checking against the COBie data and we’re finding there is inconsistencies between the information that’s in the COBie, or the model, and the information that’s in the schedules. So these door schedules for example, there is a mismatch between what’s in the model and what’s in the schedules. And also the model has only got 20 different door types whereas the door manufacturers, there was 110 different door types. Because each door type had got a variant because it’s got different ironmongery.
And you can understand why that’s happened because the BIM model is largely a design model rather than incorporating the detailed manufacturer’s information, which of course it should do. So the as-built model should reflect what has obviously actually installed, but the information that we receive from that is not easy to interpret. So the team have written agents that will then open up this PDF report and on page 33 of this report we’ve actually got what’s almost unintelligible, but the agent has then brought it back out into an Excel format.
So we’ve now gone from the PDF to an Excel format and we’ve also done the same with the other schedules, and again we’ve produced another Excel format which we can then federate into a common set of schedules. And what we can also do is compare that with what’s on the drawings. So in this case you’ll see here that there’s a door there and that nomenclature suggests or states that it’s on the second floor, whereas actually this is a first floor drawing. And then we identified that that same door reference is on the fourth floor and on the fifth floor.
Basically they’ve just copied the drawings up, but that’s a digital defect that could be a problem down the line because all of the information about that door isn’t then correct. So if you’re interested in that type of automated auditing of health & safety files and O&Ms to pick up when there is any inconsistencies, if you add ‘validation’ in to the chat we can cover that. And if you’ve got any projects, one of the FM companies have now given us information on all of the variations that they’ve had carried out by specialist subcontractors so that we can analyse those. And the results of that can then be fed back to the contractors so that not only can they get it right, but also they can learn next time that somebody is checking and then we can move from there.
Richard, could we just look at whether we’ve got people that are interested in that from an approach perspective? I think we’ve got people that are signed up, could we discuss that, do we think that’s a good thing to be doing?
RICHARD FREER In terms of the validation, there’s a number of people that have put their names down to get involved in the validation. Has anybody got any comments or queries on that? Is it clear exactly what he is talking about?
FRANCO PITTONI I'm going to repeat something that I've already said in other sessions. The first step of the validation is making sure that the documents or the drawings that are part of the list or schedule submitted actually have been submitted. In my opinion there is a step before that which is there must be a way of making sure that the list is comprehensive of all of the assets. Because there is a difference between saying there are 1,000 drawings in the list and 1,000 drawings exist.
But my question, and this is coming from my background in project management, is to verify that the list is fully comprehensive of the 100% of the assets that form part of the project.
GEORGE STEVENSON I’m pleased that you’ve said that because that’s exactly where we’re going with this. We’ve got a partnership with Morta, they’re involved by a lot of contractors and some designers as well in compiling the information delivery plan. So the information delivery plan is the starting point for that and that is normally an interpretation of the information requirements that are set out at the beginning of the project. So if that’s the case that can become a cornerstone against which information is then tested.
STEVEN NAYLOR I’m from the Health & Safety Executive, based in HSE’s science division. Just to say that we’ve been doing some of our own work, very similar to what George has described. Essentially looking to see how we can potentially link a project’s work schedule with its design risk registers, its construction risk registers, rams documents, and basically use the work schedule as a way of delving into the detail around how health & safety risks are being managed over the life cycle.
So the idea is once a project reaches RIBA stage 4 and work programmes are being created, what are the practicalities of linking all of these different registers and data sets together. And one of the things that we’ve been really interested in is how we can potentially use ISO19650 Part 6 as a way of doing that. Because what we’ve found is that all of the different registers that we’ve been trying to work with they use different nomenclature, different terminology, that they’re referring to risks in different ways and task descriptions.
So our key learning is that if projects were more aligned with the ISO standard then we kind of feel that the practicalities of doing this would be essentially a lot easier to do. And then the utility of having that knowledge resource would be there for projects. But really interested in what you’re doing in this space as well because it does sound like you’re trying to do similar things from a challenge perspective.
RICHARD FREER Absolutely, you’re input into that would be very helpful. Are you going to come along to the validation?
STEVEN NAYLOR Yeah, I’ve put my name forward, it would be really good to find out what you’ve been doing in that space.
GEORGE STEVENSON We’re actually more aligned, Steven, than you might know because Gordon Crick and Chris Lucas have been very actively involved in what we’ve been doing over the last 2 or 3 years. So one of the things that we’ve been doing is producing with Gordon is we produced a standardised health & safety file structure and methodology that we can then apply. So a lot of your thinking, we’re coming from the same point.
Jac Parry, I think you’re now with Steve. Jac’s operation they did lots of projects, so I imagine that generating lots of projects, some relatively small and some much bigger, you produce health & safety files for those. And that information needs to then get federated. Jac used to work for Mitie, a lot of their work would then presumably then have to get federated back into a more coordinated set of information, is that right?
JAC PERRY Yeah, correct, and it was always a…with the sheer volume of different manuals and different asset manuals and everything that we had to collate, which is always the sticky point by the end of a project. And, unfortunately, most of the time it got either pushed to one side, but surely because of the amount of work involved in doing it, and there was just no headroom maybe in the project to finish it.
GEORGE STEVENSON So from that perspective I’d very much like you to be part of that validation group if you can.I think we’ve got a really strong team here. The other thing, Steven, that Franco is doing with this process, what we’re trying to do, again tying it together with tools like Morta, is identify what tasks need to be done. Because if we can identify what the activities are and what tasks and needed to be done and who is producing what information and why and who is responsible for it, that’s a structured methodology of working.
So we’re running a session on the 12th of May on tasks, so if you’re involved in doing task information delivery plans or generally looking at who is going to produce the golden thread of information, if you put ‘tasks’ in the chat we’ll invite you to that. We’ve already got 50 or 60 people that have signed up for tasks.
JONATHAN HALLAM From my previous role with Jac, but on the other side of the table from a client owner from the Ministry of Defence. And it was always trying to get, even if there was lots of small projects and things coming over it was how to readily and easily get them over into the owner’s asset information model and keeping that up to date. Which I know George, we’ve spoken separately. And even though I keep in touch with some of the people who have now moved on it’s still not there. And I think it would be the same for any big landowner or big owner who is having contracting body do lots of different small projects, trying to get it into their own system.
ANDY GREEN Just to let you know that the FMS002 standard is about to be amended and will provide better guidance on how to land this into a standard asset register. So that’s due to come out to market by the end of next month. Really because we all want to bring the data across, but if we don’t have a standard way of presenting the location asset registers then it comes across and doesn’t get used as we would like it to be. But that’s really an important thing to be aware of to share that with the group.
TIM AIKIN I’d love to take credit for this slide presentation but you’ll see at the end why I won’t. What we’re trying to do in the AI & Data Engineering workstream is I’m trying to import education so that we know what these things are and we’re also getting people to present stuff. Now because of the way that this presentation has been put together that’s not necessarily going to credit people with certain stuff, but I’ll credit them now.
We had a presentation from somebody who was working with PFI hospitals and using object recognition in photographs and linking them back to HTMs which is the health technical memorandums. And there was another person who did a presentation of stuff that they’ve got where they’ve got a specifically domain trained object who are doing documents which dovetailed quite nicely into what I was talking about as well. So on with what we’re trying to achieve.
We’re trying to move this beyond the people’s first introduction to it is not just a chatbot anymore. And we want to understand and evaluate and we want to apply what we’re learning. So this is very much a sort of forum in which people are asked to present things and are asked for their comments and experiences, but also we’re trying to inject some real world examples. Sometimes it’s new technologies, so I’ll just try to knock up something that’s relevant to show how things are going to change our world, and it is changing very rapidly.
So the first things that we need to understand is that we know that it’s not just what we can use day to day and ask about, check people’s facts and things like that or get it to write a nice rhyming poem for a birthday greeting or whatever, it’s gone beyond that. The key thing is does it really understand our domain? And the thing is it doesn’t until we start to connect it to datasets. So data information like building models, documents like O&Ms, health & safety files, building log books, operational data information, IoT systems. You name it, if we can plug it into there what can we do with it?
And obviously the big questions for society and particularly the workplace is is it replacing us? No, it’s not. It’s doing those things that we find take us time, sometimes a day, and we can’t be consistently trying to do something. If you ask somebody to look at 5 files that’s a lot easier than asking them to look at 500. So it’s important that we actually understand, we are the masters and the orchestrators, not the AIs themselves. Although in my recent experience I’ve started to wonder how far it’s going, I’ll explain that tomorrow.
So the shift has moved, it’s gone from answering questions to completing tasks. So the stack we’ve been working with, what bits are they. We have AI agents, we’ve got a thing called MCP Connectors which sounds complicated but it isn’t.And we’ve got Data and Systems, that sources those databases and different types of databases. And then we’ve got Skills and Skills are where we give the instructions and we can embed certain rules, so it doesn’t go off down to rabbit holes. We can conduct and constrain what its remit is.
And that’s why we call them agents, they’re like secret agents, this is your mission, should you fail we’ll deny all liability. The agents are basically doing things that we have instructed it do. So what have we covered in the sessions? The first three sessions we’ve already covered how agents work, done some real world examples such as the product data agent where I created two different agents in two different systems. And we’ve done what George mentioned before, the audit strategy agent which is looking at are these O&Ms compliant, are these health & safety files complete.
And giving us meaningful reports in a consistent manner so that we can then make recommendations and judgements as to whether we accept information, require information to be completed, identify the gaps or accept the risks. And that risks is a critical one, that’s what insurers are going to continually ask us to prove to mitigate the risks. Tomorrow we’ll be covering, I’ve got a quick presentation on what MCPs are and try and take away the mystique of some of these acronyms that people like to confuse the world with.
And we’ll also be looking at understanding what vector databases are. Why are they different to relational databases and spreadsheets and the like? The real world examples we’ve done, these are tackling problems that we as recipients of information are constantly having to deal with. And sometimes the data can be as old as 20 years, so we need quick ways of actually understanding without having to open up every single document. It’s just not practical when you’re dealing with on average between 10,000 and 20,000 documents.
The one I mentioned before which was the product data capture which was the agent where I created two versions of it. And then the compliance audit. So we’re trying to see here that there are numerous examples we can start to come up with. So we really have got a generic pattern that’s working. We’ve to repetitiveness, we’ve got multiple data sources, we need structured output which is currently done manually. Those are the things that we’re tackling. Things like the COBie complete, what do you do to triage the documents?
How this effects into soft landings, how do we actually start to make all of that more efficiently happening? And we then need to benchmark these against different projects as to what’s happening. So the next steps that we’re going to get involved with in the workstream. These are a few, I’m sure there’s many more and what we’d like is more people to get involved and to join a session. It is a workstream, it’s not a lecture and the more that we can have people bring their problems and their solutions, the better it is going to be.
The interesting thing that I found about this world of AI, I mentioned to George that suddenly one company had all of their accounts suspended because they have apparently infringed something. And I said one of the important things is that a lot of this stuff is interoperable. So if you’re using Anthropic which is Claude or OpenAI which is Codex and Gemini is Google, anyone of those services you can interchange between the two. And in fact what I’m doing is using one agent to actually check the work of another agent which is critiquing software which is a common thing to do.
GEORGE STEVENSON I think the point is that what we’ve got is a dedicated workstream which you can join. And we’ve been meeting for the last 6-8 weeks and we’ve been learning from each other. The important thing here is that if you’re involved in this, either as a potential consumer of it or in fact a creator of agents, then come along because we can share experiences and learn.
TIM AIKIN Why I don’t want to take credit for this presentation is because it took 10-15 minutes for Claude to produce it and the only thing that I’ve done on it is put the builds. It’s built upon everything that I’ve been doing though. The instructions I gave it, I said I need a presentation for BIM In Asset Management April monthly forum. What I’d like to cover is how we are in the AI & Data Engineering workstream covering what AI is and how it can benefit and provide real world examples, that was the prompt.
RICHARD FREER Tim, did you refer it to the recordings? I’ve done that, actually referred it to the recordings of our meetings.
TIM AIKIN I have done that on one before and create a presentation from it. Not on the BIM In Asset Management recordings, but on another recording that we had. So we had to critique over a presentation and there was a series of action points that came off. And then obviously when you store recordings on Teams you’ve then got to transfer it to another thing and I gave it the transcript and I said can you create me slide presentations to help resolve some of the problems.
And it couldn’t take the slide deck that we’d used for that presentation and while I was trying to work that one out it actually wrote the other slide deck based upon the transcript. But I’ve also used AI to create an application and it creates a working prototype and creates me a slide deck for promotion. And then I’ve offhanded it to the coding agents so it’s now coding it up.
RICHARD FREER I’ve used it just with a referral to our BIMinAM website page that’s got the recordings for say the design group and I’ve told it go and look at the videos for the last three meetings and it’s come up with a summary of the three which is pretty amazing.
GEORGE STEVENSON The important thing here as far as this community is concerned isn’t that we’ve got LLMs that can do some very clever stuff. The main thing that we’ve got here are the people that have got the knowledge to understand what good looks like. Because as Tim will tell you there’s lots of things that we ask AI to do and the first time it comes back needs tweaking or it’s not quite right. And therefore what’s really important for everybody in this group to understand is that they’ve got a role to play, even if they’re not technical. They’ve got a role to play because they know what good looks like and we can then use that to refine.
TIM AIKIN And it is an iterative thing. On a project that we’ve got involved with UCL we got a set of IFC files and I started running, because I had built an MCP connector to it it read all of the files and I said is there anything useful in these models? That was the first prompt and it basically went off and gave me a whole series of things. And then I asked it more questions and then I interjected and said such and such. And then I got it to do a slide presentation and then right at the end I said how long did it take to do this? And it said it would have taken being sure had you not interjected with this…and I sent to the stuff off to people and I said I can’t take credit for this.
RICHARD FREER Without more ado, something we’ve been a bit remiss in not including and giving attention to which is water quality.
GEORGE STEVENSON One of the key things that we’re trying to do here is make sure that we’ve got coordinated asset information that can be used for lots of different purposes. So obviously a big driver in what we’ve been doing has been around fire safety, but we’ve also been doing work around air quality, in particular around damp and mould. So the purpose of this is to look at this from a perspective of what are the disciplines and what are the risks that could be being addressed by having coordinated asset information.
So what I’ve asked Gordon to do is come at this from a point of view of water quality and in different ways. He’s a specialist in heating and cooling systems, but there’s other ways in which this can be approached. So we’re running a dedicated workshop on the 22nd of May, we’re going to go into some detail here that Gordon is just giving us a snapshot of.
GORDON PRINGLE Just a couple of things, we talked about insurance, we’ve talked about risk and validation, we’ve talked about task, we’ve talked about AI and Tim just mentioned terms. Sadly I am going to probably introduce you to another one in these slides. The slide in front of you shows a level of water quality that we would wish you to have and everyone of your heating and cooling installations throughout all of your buildings, whether it be residential, healthcare, education. You usually want that crystal clear debris, particulate free water in every single installation that you potentially have.
And you can achieve that but you actually need to have some form of monitoring that process to make sure that you are achieving that. What you do not want is despite traditional standards of water sampling, which may be very sporadic, we are probably common observers of some of the failures on water quality that you’re seeing onscreen at the moment. Or indeed as we’ve experienced with some of our brand new equipment that’s gone in to existing installations where the water quality within months has depleted the performance of an asset within months of it going in to a new installation. And I don’t think that we’re alone in this aspect.
You can see this is an installation of a column radiator in a project in London. You’ve effectively got a brand new radiator because you can determine by the size and the condition of the steel work, the upper chamber of the radiator. But you see that you’ve got a tidemark and because you’ve got the presence of that air you’ve got complete degradation of this column radiator. This would be in a residential development, probably on a heat network where new regulation HN tasks now is a regulatory requirement that you have condition based monitoring of water quality.
And that’s to ensure as per the Building Safety Act and as per the golden thread, that you do not have water penetration and failure in residential blocks inside your homes and your daughter’s and son’s homes as they jump on the property ladder, because that’s the new carbon reduction policy of the UK government. But effectively you’ve also got the impact on poor water quality on reciprocating plants such as pumps, valves, seals. If your water is not clear and free of particulate this is the damage that’s going to achieve into that process.
So water monitoring for longevity, water monitoring should not be considered simply as a retrofit once a system has gone awry, but properly included as part of an initial design. So in actual fact we’ve been working very hard since the introduction of the product in 2017 to get it in to UK guidance. It is now as best practice in numerous UK guidance, BSRIA and CIBSE guidance. It’s now concluded in the draft of HN tasks, it’s now an absolute requirement in every heat network, both existing and future heat networks in the UK. And effectively what we’re doing is, I mentioned an acronym before and I mentioned topics that were discussed before.
With regards to insurance we spoke to the Association of British Insurers (ABI) many, many years ago when we picked up one of their documents that water loss was the biggest risk that the insurance industry had to cover. So pipework penetration, product failure where mass heating and cooling systems are discharging their water very quickly into your buildings is a massive insurance risk. So managing risk and getting an indication of the risk being maintained and validated and tasks to be followed would simply imply that what you want to have is a new acronym, Average Yearly Corrosion Rate.
And if you were getting that determined and predicted to you every 7 hours every single day, a thousand datasets a year, to tell you that you’re either in a good situation, an average situation or a bad situation. And if that alarm goes to the BMS head end and an engineering interface is required you’ve basically got a measurable indicator of system health and efficiency and should be systematically delivered to clients in an non-intrusive testing and monitoring technique that basically you would be happy with and you would be content.
So you want to be down here with all of those datasets, and if you’re up here at handover you don’t let the contractor go away until he gets you down here. And if he doesn’t know how to get you down there dismiss him because he doesn’t know what he’s doing. And thats’ half of the problem with regards to water quality in closed sealed heating and cooling systems. So that’s probably as much as I want to show you just now because it gives you a bit of a taster as to what we’re trying to achieve.
We’re trying to achieve to give you a repetitive dataset on water quality in heating and cooling systems continuously that allows you to take comfort from that. And if that’s mitigating an insurance risk then surely that must be reducing your premium on those systems. Because if their biggest loss is water loss and that involves heating and cooling systems, then what are they going to do about it?
RICHARD FREER So if you’re interested in coming along to Gordon going in to a lot more detail on that put ‘water’ in the chat.
GORDON PRINGLE Water in closed seal systems is your energy around your buildings. So you’re basically heating and cooling all of your buildings generally with water.
GEORGE STEVENSON Something else on the water side of things. We won’t just be covering in that workshop the heating systems, we’re also going to be including the other side of things which is things like Legionella. So we’ve got another partner who we’ve been working with for some years now called Accoutro and they’ve got a solution which the FM contractors may be able to use to actually cut down on the amount of specialist consultancy or specialist contractors that they bring in to do water treatment.
Their solution is called Waterguard and that certainly looks as though it’s something that complements well what Gordon is providing, but I think it’s a more general water management and compliance solution. I don’t know the details of that, but it’s being implemented at the moment at quite a few major contracts. A lot of the big office buildings like 22 Bishopsgate in London are using that. And it’s also being used to identify how the length of some of the lifecycle risks of water maybe damaging things.
GORDON PRINGLE Just a point on that, we have got large national asset management companies that are basically now introducing monitoring on systems ahead of procurement and purchase of buildings. Because the clients are effectively saying this is an unmanaged risk, so we want to know what we’re buying. The water system that’s in that building, what is it doing to the material that’s in that building? I think can we take a view on the sustainability and longevity of that system.
GEORGE STEVENSON Yeah, they were telling me for example, there is a building that was only finished 6 years ago that maybe has got some serious issues where the heating system I would think would normally be expected to last for at least 10 years, probably longer. And already there is some issues there because of just reaction between the chemicals are having on things. So that’s a really interesting workstream that is looking not so much at safety risk…Well, I suppose certainly in the Legionella sense it is very much looking at safety risk, but also at operational costs.
GORDON PRINGLE You’ve got 82 degrees water above your head, it’s a bit of a safety risk if it penetrates above your head I’ll tell you!
MARK AUSTIN Just one question, with the new RICS standard around the use of AI and engineering, are we aligning all of the work that’s being done here to those requirements with the professional bodies in mind?
GEORGE STEVENSON No, I’m not aware of what those are, to be honest with you.
MARK AUSTIN Basically RICS as of March issued a professional standard for all chartered surveyors and chartered surveying companies around the use of AI and how you put your professional judgement within it and how you act as professional people. And obviously RICS have a huge involvement in this sort of sector. So I was just interested to understand, to try to get to grips with it myself.
GEORGE STEVENSON No absolutely, as you know we’re all volunteers and we’re here to, if you’ve got some recommendations as to what would be good for us to incorporate I’m happy for you to….
MARK AUSTIN I’ll drop the standard in the chat and I said I’ll join the group.
ALAN OLIVER Another great forum, just one observation. George, you mentioned about the works package, whats missing. Well, clearly you need to know what’s missing to know what’s missing and that’s one of the issues. And as-built buildings, getting it right is the key to asset management. And if you look at the Midlands Metropolitan Hospital, although it’s a great asset register it doesn’t actually tell you, for example, whether all of the fire doors have been installed correctly. And there’s got to be a join up, it’s got to be what’s happened front of house with the as-built.
MARGARET SACKEY Just picking up on my RICS colleague comments because on behalf of the Institution of Civil Engineers I lead the safety expert panel and also I was very much involved with Gordon Crick and Chris Lucas in terms of the previous 1192 and now 19650 Part 6. And in terms of AI I’m really interested in if RICS have issued a circular to its members I am interested in understanding that. But also you’ve got Nima which hopefully have got something to say about AI and responsibility and the rules and this sort of thing.
TOM OULTON I’m from Constructing Excellence and all things digital. Nima have released a statement on BIM and it’s in conjunction with GIG. I’ve got a copy of it, I’ll put it in the chat here.
CHAT
Paul Toner…Exactly Ian Upton, all readers should be banned and people actually pay attention to the forum interactions, and the uploaded video recording. As it seems like that they are not really invested in this process
Phil Thompson….Michael Chambers how do insurers propose to trace products via the golden thread, that have been delivered to a project as product and not an asset?
Phull, Pav Singh….Do you have industry guidance we can read on insurance / liability
Pierre Venter….PI requires competence. measured and certified to address the risk
Dean Dixon – SmartStruct….could the golden thread be managed via workflows within a CDE platform that insurers could login to manage any arbitration?
Mcgeachie, Natasha….Within Skanska we conduct something we call ITP testing. and add this into the project information
Pierre Venter….The documents to define the services need to be really robust. that’s a challenge.
But are we being paid for it
Tom Oulton….Speaking to insurer's a few years ago, they asked "Will BIM make the building cheaper to rebuild? If not,, the insurance remains the same!" Q - was I talking to numpties?
Grant Bezuidenhout….Surely it is about robust requirements within the contract scopes which are then managed with suitable governance and assurance, linked to BSA requirements and a full audit trail. I don't think it is complicated at all, it is information management.
David…..Apologies I have to catch a train. I want to bring to your attention that we have developed a new variation of our HOCL solution for specific application on mold. We have designed a solution with a surfactant combined with HOCl which is gel based so you can apply the solution directly onto the problem areas without any fun off. The surfactant will remove staining and the HOCL will kill the spores. In addition we are testing our solution with Electrostatic sprayers to look at effective and efficient application to the surface areas. A pre calibrated label/Test strip is being developed post the application as proof of a safe application to the surface and the surface clean and safe. Happy to discuss this further and host a call if there is interest.
Pierre Venter…..@George. the Golden thread is broken in real terms when the model data is shared in procurement, supply and installation
the chain of decisions from the initial decision to as-constructed and delivered, is not continuous.
the argument is levelled that the design model, is not consumed by the unsophisticated user, but the GC has the capacity
the standards, competence and capability is dismal across the role players , as it’s an expense. insurance would change that, except that’s like adding lawyers to cover yourself
Tom Oulton….The Validation exercise is integral to the As Built definition project
Fabienne Kentish….Good morning everyone. By way of introduction, I am Head of Business Development at CHPK Group, a multidisciplinary construction and engineering consultancy spanning Building Consultancy, Building Control, Building Safety, Façade Design & Engineering, Fire Engineering and MEP Services……https://chpk.co.uk/ I heard someone mention Gateway Three and new BSR Rules - if anyone is struggling to get their head round it, which I don't blame them, my head of Building Control is a walking encyclopaedia on the matter and regularly speaks at RICS events. He's always happy to have a chat, or host a CPD. The team works very closely with architects, developers, designers and contractors navigate Building Regulations to deliver projects that are safe, compliant, and ready for approval. If anyone would like to discuss any of our other services, please feel to reach out on email fabienne.kentish@chpk.co.uk
Emeka Efe Osaji CEng FCIBSE….Hi everyone. I am so sorry, but I will have to leave at 12pm for another engagement. Please accept my apologies. Many thanks for another brilliant meeting. The presentation and discussions have been great. Wishing everyone a fantastic week.
Dan Hoogeveen…Steven Naylor Would be great to connect outside of this meeting in relation to BSR application validation if youre in anyway involved in this.
Mark Austin….Happy to share the work we have been doing on data discovery around project information in the infrastructure sector that already exists at the validation session
Jac Parry….Tim Aikin / George Stevenson - We're working with a delivery partner to build an entire AI overlay platform to deliver just this, and would welcome further discussions and concept development.
COX, Melanie Jacobsen (NHS ENGLAND)….. So pleased to see Water on the agenda - for the healthcare estate safety fundamentally relies on water and air quality, in addition to the BSA focus on Fire and Structure.
Gary Watts….Hyde Park One. Corrosion quoted as the cause of Building Safety damage. Any thoughts on why and what should have taken place to mitigate.
Margaret Sackey….Hello RICS - Mark Austin - I represent ICE in terms of Building Safety and ISO 19650 - George and I are at RICS tomorrow for the Construction Industry Task Force Forum - the Data Management Workstream would benefit from your RICS Guide. Can I contact you - My email is. Marsackey@gmail.com. Many Thanks
Margaret Sackey….Hi BIM 4 Asset Management Team - - please can the links provided by Mark Austin and NIMA be circulated to attendees. We really need to collaborate on this issue. Thank you
Structured Academic Summary: BIM in Asset Management April Monthly Forum (29-04-2026)
This document summarises a multidisciplinary forum focused on advancing digital asset information management within the built environment, particularly in the context of the Building Safety Act, Golden Thread requirements, and Gateway 3 compliance.
The forum emphasises a paradigm shift from traditional BIM as 3D modelling toward BIM as structured, lifecycle information management. A central theme is that incomplete or missing information constitutes both an operational and statutory risk, increasingly framed as a “digital defect.” This concept extends contractual accountability to information delivery, potentially linking it to insurance exposure and claims.
A significant contribution relates to insurance integration. Early engagement with risk advisors, structured data presentation aligned with insurer expectations, and narrative-rich submissions are identified as critical to achieving favourable insurance outcomes. Poor data quality leads to exclusions, inflated premiums, or limited coverage, whereas structured and validated data enables more precise underwriting.
The discussion highlights systemic issues in handover processes, particularly fragmented and unusable datasets (e.g., extensive fire stopping reports). Solutions include transforming document-based information into spatially coordinated asset information models, integrating asset data, location, and risk overlays into unified environments.
Validation emerges as a core methodology, comprising three stages: verification of information completeness, assessment of content quality, and reporting of deficiencies. Automation through AI agents enables scalable auditing of large datasets (10,000–20,000 documents), identifying inconsistencies between BIM models, COBie data, and O&M documentation.
Cross-industry insights, particularly from HAZOP methodologies, demonstrate alignment between risk identification, mitigation, and lifecycle data requirements. Integration with standards such as ISO 19650, ISO 8000, and ISO 55000 is essential for consistency and interoperability.
Emerging technologies, including AI agents, vector databases, and MCP connectors, are reframing workflows from manual review to automated task execution. However, domain expertise remains critical to define “what good looks like” and validate outputs.
Additional risk domains, such as water quality in heating and cooling systems, are introduced, highlighting the importance of continuous monitoring and its implications for safety, asset longevity, and insurance risk reduction.
Overall, the forum reinforces that coordinated, validated, and structured asset information is fundamental to regulatory compliance, risk management, and operational efficiency.
Key Insights
- Information gaps are emerging as ‘digital defects’ with contractual and insurance implications.
- Golden Thread delivery must be evidence-based, continuous, and validated at Gateway 3.
- Insurance outcomes are directly influenced by data quality and early engagement with risk advisors.
- Current handover practices produce unusable datasets despite technical compliance.
- AI-enabled validation can transform large-scale document auditing and compliance verification.
- Fragmentation between BIM models, COBie, and O&M data is a systemic industry issue.
- Standardisation (ISO 19650, ISO 8000, ISO 55000) is critical for interoperability and scalability.
- Domain expertise remains essential to guide and validate AI outputs.
- Risk methodologies such as HAZOP align well with building lifecycle information requirements.
- Operational risks (e.g., water quality) are increasingly linked to data and monitoring strategies.
Action Points
- Implement structured validation processes for all handover information (completeness, quality, reporting).
- Define and enforce information delivery plans as contractual baselines.
- Engage insurance advisors early and align data outputs with insurer requirements.
- Adopt spatially coordinated asset information models integrating assets, risks, and documentation.
- Deploy AI agents to automate document auditing, data extraction, and compliance checking.
- Standardise data structures and nomenclature using ISO frameworks.
- Establish governance to ensure continuous Golden Thread integrity across lifecycle stages.
- Integrate risk methodologies (e.g., HAZOP, bowtie) into asset information workflows.
- Introduce continuous monitoring for critical systems (e.g., water quality) to reduce operational and insurance risk.
- Develop cross-disciplinary collaboration between designers, contractors, FM teams, and insurers.