Accelerating Design: Civil 3D in the Cloud and a Journey to Global Deployment

GLENN LLOYD: Welcome to our presentation, Accelerating Design, Civil 3D in the Cloud, a Journey to a Global Deployment, Buro Happold's story. So today, we're going to cover defining what our problem was, our road to the solution that we came up with, the solution, what it was, and some key takeaways from that process.

DES KINLOUGH: Before we get started, let's introduce ourselves. My name's Des Kinlough. I'm an Associate Director at Buro Happold, and I lead the BIM, and I'm the BIM Lead for the UK side of the business.

GLENN LLOYD: And my name's Glenn Lloyd. I'm an associate at Buro Happold. I lead our Cities BIM and Technical teams.

DES KINLOUGH: So before we get started, we thought we'd give you an introduction to the company. We're a global company spread over 31 offices and four regions. We have over 3,000 staff. And we pride ourselves on being pioneers in digital design.

We cover all the major disciplines with our core disciplines being civil, structure, and MEP design. But we have an additional 28 other disciplines in the engineering field. And one big thing about what we'll go through today is we pride ourselves as operating in one collaborative community.

GLENN LLOYD: And a little bit about the Cities team that I head up. So we're about 300 technical staff across 10 of our offices. We offer design services in all aspects of civil engineering, but with a particular focus on roads, drainage, earthworks, and underground utility design. We've got about 150 civil design software users, and that primary tool set that we use is Civil 3D, AutoCAD, Navisworks, Infraworks, InfoDrainage, and then a variety of GIS and programming softwares, FMI and Dynamo, for example.

DES KINLOUGH: So just to touch on our one collaboration space, we've got, like I say, 29 offices utilizing the cloud. We're at over 840 projects stored on the cloud, and we've got 4,500 models inside those projects. And we have just over a quarter million syncs. So we're extensive users of the cloud, and this is why our journey into Civil 3D utilizing the cloud was so important.

So like I say, we work in one collaboration space across the world. So we share our projects, which are quite big, major projects, and we have to share them across the world. So just to give you-- we harness all the data that we have in the cloud from our Buildings team. We've been utilizing that for three years.

But as you can see on this diagram on our major projects, we're sharing between up to 21 offices on one project down to 10. So utilizing the cloud in our company is crucial to our success.

Our Buildings team adopted the cloud solution mainly because Autodesk offered that solution for the Revit side of the team. But it was really important that our Infrastructure and Cities team also had that ability to adopt the cloud. And that's why we started talking to Autodesk about how we could start adopting the cloud for our infrastructure team. Now Glenn will just take you through and define the problem.

GLENN LLOYD: Thanks, Des. Yes. So where we were around 2021, 2022 when the world kind of changed a bit was we were really happy with our processes. We knew how we worked and we'd come up with quite a good system, but we were working on local servers spread out across the world.

And a project would be set up on a single server, and then it would be basically wherever the project lead was. And so what that meant was if they were a project team member not from that office, they were having to work from home or remotely in general via VPN or virtual desktop, remote desktops, or sometimes worst case, transferring files backwards and forwards. And then obviously when the pandemic hit and everyone was working from home, that meant everyone, even people in the local office, were working that way. So it wasn't great.

A few years prior, we'd had a really big focus on moving all our project setup from maybe 3D and 2D running concurrently to having a 3D, Civil 3D model as our single source of truth that generated everything from our xrefs for our sheets through to our collaboration models for however and whatever platform we were collaborating on. So that was-- whether it was Navisworks or similar.

That looks simple enough, but when we expanded out to a full project-- and I think this was basically not even that big of a project-- we have a lot of data shortcuts in the green and a lot of xrefs going on in the blue to make that system work. So while we have our design models that are our single sources of truth highlighted in yellow there, there's a lot of referencing going on. And therefore, anything that slows that referencing down was very problematic for us. So we liked the system. We just didn't like how it worked.

So we set about defining what the challenges we were facing were, and then we started working towards a solution. And Des and I will go through those challenges now.

DES KINLOUGH: Yeah. Thanks, Glenn. So the first challenge-- and I think everyone who's listening today would have the similar challenge-- is there's downtime due to opening and referencing and saving and printing the files. We turned our IT system inside out, really, especially during the COVID time and the work from home environment, which is still quite prevalent today. So trying to access files from home through a network was really slow. So time-wise, it was a big challenge for us.

GLENN LLOYD: And the next of the challenges were collaboration. So as Des mentioned earlier, we want to be a collaborative office. But what happens is when it was too slow to collaborate, you would quickly default back to working in a single office. And then that meant what we were trying to achieve wasn't possible. And so the best practice actually became, well, the project's set up in this office, so let's pull a project team from this office as well, and that's not what we were trying to achieve.

DES KINLOUGH: We started to look for solutions from the work from home environment. We were looking at other cloud solutions, but we were also looking at some virtual machines that we would have to purchase and then some remote machines. And this was a stopgap investment that helped us, but obviously there was additional hardware that we had to purchase. And it did improve things, but it didn't solve the problem wholly. So it still required additional people needed on teams for simple tasks like printing on issue day.

GLENN LLOYD: It was also a big risk to the projects themselves with users sometimes having to run projects offline or off the server. Backup processes, et cetera, weren't being followed. And it meant sometimes reviewing couldn't be done in the best way. For example, if everyone was in a very slow environment, drawings would just have to be PDF'd and sent across for review rather than the reviewer having access to the model and sometimes even not the most up-to-date data. So that was very risky to the project as well.

DES KINLOUGH: Finally, and most importantly, really, the well-being of the staff is affected. The project pressures, the deadline days were turning into absolute nightmares and affecting the pride and work that you didn't have time to finish off that drawing, that model to the level of detail you wanted.

And there was time spent when you left the office that you felt you were being unproductive, and nobody wants to feel like that. So it was a big-- a big plus for us if we could change that around for the well-being of our staff.

GLENN LLOYD: So these were the challenges we defined. And so we then set about looking for a solution. And our first port of call, as our buildings team was already using BIM 360 at that time, was to engage with Autodesk. And Des and I will now take you through that process.

DES KINLOUGH: Thanks, Glenn. So like Glenn said, we had a good relationship with Autodesk and the technical teams. And we were asking, when was this solution going to be ready? And they came to us, funny enough, around Autodesk two years ago. We met with one of the technical team and arranged some follow-ups when we got back to our offices.

And we started-- they advised us that now there was a chance that it could be compatible, and they were starting to move in the right direction. So this was a big plus for us.

So Buro Happold tentatively-- Glenn tentatively tested some files and confirmed enough on that trial to test on a project in tandem. Not live, but in tandem with a network project. And the critical findings was it was easier to move a project off the cloud than move it on. So that was good, and that was quite a big bonus to us. And it gave us some confidence to test in a low-risk environment.

GLENN LLOYD: So about mid 2022, we selected a project. It was at the early stages of a project we knew would only have five to 10 users, and they'd be across our UK and Mumbai offices. So a good project to test the cloud data on. It was going to be a lot of underground utilities and drainage and a little bit of earthworks, which meant we knew there'd be a lot of data shortcuts. So if there were benefits to be found, we knew we would find them on this project.

And our initial feedback was that it was good. When we had this-- when it was working well, the speed was amazing, especially synchronizing and opening. But we did start to find some issues, particularly around files not syncing, not locking or unlocking consistently.

And then kind of a little bit harder to diagnose was data shortcuts not syncing or locking and unlocking correctly. And then the worst we found was what we described as black box errors where something wasn't working, but there was no indication that it wasn't working. So we didn't know why or we didn't know how to fix it.

Unfortunately, as we brought more and more users into the project, these issues started to compound. And in the end, the decision was made to pull the project back off the system. It wasn't quite ready for a project that big. And we moved it onto our own servers, and went back to our old way of working on that project.

DES KINLOUGH: So although things didn't work out, we still had confidence that we were heading in the right direction. And one of the main reasons we had that confidence was that we were having those discussions with this team, and we were having engagement with all the key members of the ACC team and Civil 3D teams. So we kept the projects running locally on our network servers, but we still use BIM 360 and ACC to store files and share files with other project team members.

GLENN LLOYD: Yeah. So fast forward about a year from the beginning of that initial trial project now, and Autodesk had advised us that Civil 3D 2023 and Desktop Connector, particularly 16.3, they'd made some big improvements and had solved a lot of those problems we'd been discussing with them. So we embarked again on another trial project. We did a quick test once again on some dummy data or some offline data, and then selected a project to trial.

This time, it was actually a large roading and public realm scheme, but we knew, again, it was going to have plenty of referencing going on. And we also knew the first phase of that project would just be three users between London and Berlin. So one of them was me. And so we hoped it would be maybe a slightly more static trial.

And what we found was all those issues that we'd been fighting on the previous project were sometimes still around, but were much less frequent. So we made the decision between project phases we'd bring the whole Civil team onto the BIM 360 project. And this time, it was much more successful.

We kept the project in the cloud. It's still in the cloud now. And we'll go through a little bit more about this project in a bit. But in the end, we had about 40-plus users on it, so it was great.

DES KINLOUGH: Like Glenn said, we could definitely see success in the Desktop Connector 16 series, so that gave us the confidence to move on as a team. Additional trial projects were added to it late 2023, and the results continue to be positive with a few minor issues. Again, we were keeping up that engagement with Autodesk so we could solve those issues.

Feedback from the users was very positive, so positive that people were wanting to move on the cloud. And like I say, ongoing discussions with Autodesk where we were providing the issues and they were giving us some best practices advice was really helpful.

So yeah, we made that decision to move all users from Civil 3D onto Civil 3D 2024 to 16.5 or later on all future projects. So you can see that within a very short space of time, we had gained that confidence to move a good number of projects over to the ACC platform.

GLENN LLOYD: So where we are now is that ACC is now our default location for storing all Civil projects. We agree to collaborate fully, whether that's on ACC or BIM 360, and whether that's on a hub that we host, a client hub, or the lead consultant's hub. Generally, we're pretty flexible on that.

We've got about 35 active major projects across eight different hubs, and about 75% of our users are now onboarded and working on those cloud projects. The next step is now ACC for storage is business as usual for us, but we're going to move to fully utilizing the other modules.

And it's worth noting that it has meant we've had to purchase some Collaborate Pro licenses, but we believe the investment's been worth it. And we'll actually talk a bit more about that a bit later.

DES KINLOUGH: So some stats that Glenn and his team ran on the trial project, which shows you some good stats. So from the point of view of users, you could see all these stats peak in November, which was a main kind of issue for that project.

But we had over 35 member users utilizing it. And you can see on the number of drawing files that were open was five-- just close to 500 at its peak. And then from the actual cloud saves, we can see that they peaked up at 5,000. And like we mentioned earlier, one of our issues were people are afraid to save. So the confidence that utilizing the cloud gave us reduced the risk of any unsaved work, which happens on any project we work on.

GLENN LLOYD: Yeah. And so just a little bit about that project. So in the end, we had 22 major models. So those are those yellow ones that we had in that diagram earlier. They were generating about 1,100 or 1,200 data shortcuts of corridors, alignments, surfaces, and pipe networks. Eventually, those split out into 900 sheets of drawings.

It was worked on across six offices, and there were 36 different users. And we defined those users as people who opened, edited, and closed a DWG file. So that's excluding all the Revit files that were also part of this package, which actually Des will talk about now.

DES KINLOUGH: Yeah. So big success from actually utilizing the cloud in a live environment. But also, we got some extra benefits. We were the lead design consultant on this project. And we had six companies working with us. And we utilized the cloud for additional benefits, such as the common data environment, which we were able to add 116 members to the team.

And then the design collaboration side of it where we were able to-- people were utilizing different file types. As you can see, seven file types from seven disciplines. And so we were able to use the Design Collaboration module.

And then finally, the visualization, probably the most powerful. In meetings, we were able to collaborate and federate all our models, and we were able to understand the design and any issues and utilize the visualization in project meetings as we were lead design consultant to steer all the companies in the right direction. So real big success on both fields. And now Glenn will take you through the solution.

GLENN LLOYD: OK. So where we got to, we deemed-- looking back at what we had done, we deemed that we think this was the solution. So what we achieved and what we were trying to achieve, I mentioned briefly that we were quite happy with our workflows. They just needed to move to a new location. And for us, that location is the Autodesk cloud. Initially, we were BIM 360, and now we're on ACC.

So yeah, just a quick sort of diagram of this is what we were trying to achieve before that-- before we started going to the cloud, which was everyone connecting through a variety of methods to a local server. But where we really were was the next slide-- yep-- which was everyone connecting through a variety of different methods. And this diagram just sort of tries to show how it could have gone wrong potentially and where it was going wrong.

And where we are now is this. Everyone's connecting via their personal laptops or their laptops to the Autodesk cloud, and that's where the project is. You can see just quickly in that diagram, another benefit is it's actually freed up a lot of our remote machines, which we did talk about. And they're now being used for other purposes or being replaced with additional users, which is great.

DES KINLOUGH: So yeah, looking back on the challenges, we thought we'd give you some feedback. Now we're over a year into utilizing the cloud in real time. So yeah, that time-- the downtime, that was a big issue for us and the work from home environment.

That significantly has improved to the point where previously, it would take 10 to 20 minutes to open a file. It's taking from one to two minutes on the opening time. Similar on save times. So that's an 80% to 90% efficiency gain. So now a simple task, which probably would be a big task, is not a problem anymore.

GLENN LLOYD: OK. And our collaboration has improved greatly. Our best practice, which we talked about before, best practice being run the project wherever the project's saved, our best practice is now, which team works best for this project? So we can pull the right people onto the right projects no matter where they are.

Project difficulties due to a geographic location, they're now resolved. Teams across offices-- and this is something we found-- is that they're not just collaborating more on projects, but they're actually collaborating more on learning and development and those kinds of things as well because you're talking to other offices more often. You're saying, oh, actually, this person has this expertise. Let's get them to do a lunch and learn or a training session on it, which has been great. And that was almost unexpected, I would say.

And then we've probably had currently about a 20% to 50% increase in collaboration, and we only see that going forward. We don't see geographic location as a determining factor on who works on what project at all anymore. So that potentially will raise to 100%.

DES KINLOUGH: So the investment, like we mentioned previously, we were moving London offices, which is one of our biggest offices, over 800 people. And we were looking to invest in some remote machines and virtual machines in these offices. That was now significantly reduced because of the work on the cloud.

That solution didn't need that extra investment, so that meant significantly less infrastructure on site at that time. The problems of setting up and saving could be now made in junior members. And it meant a 60% to 70% reduction in remote desktop machines that we'd planned to do just before this trial started.

GLENN LLOYD: Yep. And then the project risk side of things has largely disappeared, to be quite honest. There's now no excuse for a user to be running something offline, and that just means there's significantly lower risk of anything being lost.

DES KINLOUGH: And yes, like we mentioned previously, the well-being, it's not something that you can measure as much, but you measure, I suppose, by the atmosphere in the office. So like we said previously, people are now asking, actively asking for their project to be on the ACC BIM 360 platform.

So yeah, people feel like their time is being spent on the output, not watching files load. Deadline day is a lot smoother. And users are actively, like I say, asking to go on to projects. So from a well-being point of view, this has made a massive difference. And so yeah, the outputs can be people not leaving your company. It's as big as that.

GLENN LLOYD: OK. So we're quite confident that initially, a lot of our challenges and those kind of things have been solved. But we also wanted to have a look on return-- look at return on investment. And obviously, we needed to feed that back up the chain because we were asking to buy a lot more licenses.

So we sort of polled the team and we grouped everyone into two groups, either light users, which is basically an engineer, someone who needs to open a couple of files a week or maybe has one busy day where they're opening lots of files. And we estimate they save about half an hour a day if they're working away from the server, so to speak.

And then our heavy users, which are our Civil 3D technicians that are maybe opening five drawings, working in Civil 3D all day, every day, updating data, shortcuts constantly. Conservatively, we estimate they save about an hour a day. And some of them would tell us significantly more than that compared to when they're working out of the office over a VPN or a remote desktop.

And so in our company, at our current usage, we've got about 50 light users and 30 heavy users under that situation. And we currently have about 60% of our projects on the cloud. And we estimate most people are working out of the office, whether that's at home or on a different project, a project based out of their office about half the time.

So we sort put those numbers into our calculation, and then we're going to talk the numbers that we ended up with in the next slide. But you could apply those numbers to your business, or different numbers to your business if needed.

DES KINLOUGH: OK. Thanks, Glenn. We'll go into more detail now on the investment versus return. So just a reminder on the returns, we had an 80% efficiency time increase of opening files. We had an increased collaboration across all our regions, all our offices of 35%. And we had a reduced hardware investment up to 65%.

Some soft returns were more time in designing and meeting our clients, which is really important, as well as the increased well-being of our team, like we've gone into previously. It's a big return on investment.

So our numbers are based per user. And we've looked at-- we bought 100 additional licenses for our company, but it could be different for your company. It could be more, it could be less. And the cost price in the US on the Autodesk website is now $945 presently.

So like Glenn said, based on a light and heavy user return on investment, this is quite conservative, we feel. We feel that 40 hours per year per user is saved. And our charge out rate would be $150 per hour, so that's a return on investment of $6,000.

So that's quite a big return on investment. If you look at just $1,000 for a license and a $6,000 return on investment. So we feel, like we say, it's something that we've pushed across all the companies and we'd recommend for anyone that's watching this video.

GLENN LLOYD: OK. So there's also an element of training required, as there is with any change. So we thought we'd talk briefly about that. Our Phase 1, as we're calling it, just working live in the cloud, our biggest finding was that there wasn't that much training required. Obviously, there was a little bit of, hey, we save our files in a different place now, and this is how you synchronize a project, et cetera.

But generally, we covered all of that off as just a lunch and learn session, and we kept those small with just the projects. As we were trailing projects, we'd bring one project in, have a quick session on it, and then we do those just with those project users. So we weren't trying to teach 100 people the same thing at once. We were just doing it slowly. And then as we rolled those projects out, we'd do a session with each project, but what we did was we ensured that there was an experienced user assigned to each project that was going to be using the cloud.

And then we had feedback sessions. And what we really found was some of those feedback sessions became training sessions because someone would have a new tip or a trick or a way they'd worked out how to do something. And then that would actually-- instead of it being just a feedback session, they would then present how they had-- what they were doing that was different. And that we found really useful.

And then obviously, with new staff coming in, sometimes they'd never seen the Autodesk cloud system, so we've added a little section to our onboarding manual. And again, they then, once they're on a project, get a specific session on that project, particularly we find if it's being run on another consultant's hub because it means the filing system is sometimes a bit different and they have to dig a bit deeper to get into what we consider to be our part of the filing system. So we do a little bit there.

And in Phase 2, which is going to be fully utilizing Autodesk docs and Collaborate, et cetera, we have started doing that now. And there are some training modules that we're creating for those on design collaboration, publishing and consuming data, file markups and issues, and then the file rollback and recovery. So we're setting up one-hour sessions on those that will be initially presented to the group and then available as a video on how to use those additional modules once you're confident just working and saving your files in the cloud.

And then we just thought it would be important to-- we talked a lot about some of the issues we were having and quickly discuss how we solved some of those. So a really big one that took a little bit of getting used to for everyone is reminding users to allow desktop connected to sync fully before they log off for the day, especially if they're going on leave or something like that. You can't just hit your final Save, close Civil 3D, and immediately turn your computer off.

You need to give it about 10 minutes. Go and get a coffee or a drink and come back. And then check it's synchronized, and then you can turn everything off. So that's a little bit of a change in behavior.

And then in order to prevent that, it's really important to make sure you've got at least one admin on a project. And this, again, particularly when working on someone else's hub, is that an admin can unlock a file. So if a file gets locked accidentally or doesn't unlock more critically, and the actual person who locked it could be on leave or just not around that day, an admin can always do that. So we found it's really important when collaborating with other companies and working on their hubs to request at least one person from our team as an admin to our own files, obviously, and therefore can unlock files if needed.

And then when adding an existing project to ACC, initially this was very difficult, we found. But it's now much simpler. And using the reference manager and the data shortcut manager to manage your paths is a great solution, especially if they weren't set up as relative paths before.

Another one we found is particularly if making bulk or wholesale changes to files-- so say renaming a whole set of files or moving a set of files from a print setting to an outgoing file, for example, it's best to use the web interface. It happens much quicker, and it doesn't have to effectively synchronize from your computer to the web, and backwards and forwards constantly. So we find that's really helpful if you can, using that web interface to do it.

And then if all else fails, Desktop Connector/troubleshooting. So you right-click Desktop Connector and go to the troubleshooting. It's actually new, and it's really useful for finding files that might be stuck or are on your desktop, but are not synchronized correctly. And then you can work out what to do with them.

And then really, if all else fails, exiting and restarting Desktop Connector we've found really gives it a good kickstart. I'm not really sure why, and maybe someone from Autodesk can talk about this at another session or somewhere. But just restarting your computer doesn't necessarily kickstart it, but exiting Desktop Connector and then restarting it quickly does. And sometimes that's the best solution, and it just seems to kickstart those syncs, and then everything syncs back up again.

But I think it's important to say a lot of these issues we were having before, and we're having them significantly less. And we've found with every new rollout of a new version of Desktop Connector, that's less and less. And some of these, I would say, I haven't had to do in weeks. So yeah, we're really-- really finding that great.

DES KINLOUGH: Some really good tips there, Glenn. Thanks for that. So we just thought we'd touch again-- Glenn mentioned it previously on the training-- what's next for us? We've got the advantage that our Buildings team have been utilizing ACC for years. So they're using a lot of these tools, so we can tap into some of their knowledge.

So design coordination is a big one for us. We've got multi-discipline, like up to 21 disciplines. Not everyone modeling in 3D, but for bringing all those federated models together, we see that as a really big advantage inside ACC. So we're going to really push that going forward as a multi-discipline project.

So from a design collaboration point of view, this is a real big kind of time saver inside the Buildings team, and we want to really utilize this inside the Infrastructure team. It's still not there yet from Infrastructure, and there still is a bit of way to go, but it's something that we're keeping an eye on and we hope to utilize going forward.

Similar, we mentioned previously we used from a common data environment. ACC is brilliant. You can add users from multiple teams. We've just purchased an enterprise license for Collaborate, ACC Collaborate, which allows us to invite users who don't have licenses. So we can add them on for permissions. You can control permissions.

Then obviously, we can utilize all the visualization tools and the sharing tools, but also be able to just drag and drop and save files. So at the moment, we generally use our 3D files, but in the future we're probably looking to save other file formats.

One kind of area that is not really looked at is the file recovery system. Every time you overwrite a file, there's a new version that comes in, and the version is really good and easy if you want to recover files. If there's any problems, previously that would rely on going to IT departments and looking back at any archive systems. But utilizing it inside ACC is really easy, really simple, and can save a lot of time and reduce a lot of risk.

So yeah, these are some of the areas that we're looking at. There's other areas that we haven't covered on here, but we feel these four are the ones that we want to concentrate on going forward.

GLENN LLOYD: OK. So we'll take-- I'll give you a few key takeaways and our key findings, I guess. And those we think are-- and this kind of goes for any system, but definitely for this system-- developing a rollout plan, testing it on live data and accessing the support that's available, and then deploying it in a phased way. And we'll talk a bit about each of those.

DES KINLOUGH: Yeah. So let's delve into develop a rollout plan. We highly recommend when you walk out of here and you want to utilize this inside your company that you define the problem. Sit back. Like we said, what were the challenges? Probably very similar to our challenges, but you had to find that for your company and what you're trying to solve.

Sit down and write a program. Get those key milestones. Get that trial project. Get the right trial project and assess that progress. And like we showed in our journey, step back if needed. You don't have to pile through. You don't have to utilize every project in the first six months.

And identify and engage with support early. We've emphasized this a lot in the presentation today, but it's because it was really important. Engage with internal IT support teams. They're really important to support you with new rollouts, new updates. And our team were fantastic for that.

And then Autodesk support and product experts, lean on them. They were brilliant for us. And we met once a week in the kind of thick of it at the start. And all the people not mentioned here, our IT and Autodesk and ourselves worked closely together to ensure the success.

GLENN LLOYD: Yep. And then the next one that we deemed a real key is testing on live data while accessing that support. We definitely found that working with demo data or an offline project, it wasn't real world enough to really stress the system as much as it did get stressed when we were using it on live data.

We knew that with this system that stepping back was easy. We knew that we could easily-- and when I say easily, literally overnight just download the whole project off ACC, again, put it back onto our P drive, connect a few data shortcuts back up, and we were good to go the next day. So that gave us a lot of confidence.

Talking to Autodesk support, it helped us understand the issues and make decisions because it did mean sometimes they were acknowledging that that was a problem. We weren't doing anything wrong. It just wasn't working quite right yet, so that helped us make decisions. And then we had regular check-ins also with the users because we didn't want any users to be sitting back and just thinking, hey, this is normal, or I just have to deal with this now.

And then we say a smooth rollout in the end was a big collaboration between the three main parties. That was the Buro Happold infrastructure team, the wider Buro Happold team, which is IT, but also management. It needed everyone's support because there was a chance it wasn't going to work out. And then that Autodesk support team.

DES KINLOUGH: Yeah. So we really want to emphasize a phased rollout is the way to go. Start small on one or two projects as we've done. It gives you time to learn the software. Gives you time to understand the mistakes. And then before you go into a full rollout, you can apply those to a kind of larger scale and a higher number of projects.

As experts or super users emerge, ensure they're split. So don't put them all on one project. I know that sounds quite obvious, but make sure they're spread around. It helps them develop and it helps the teams that they're working on.

And then plan to grow in more projects in incremental steps so that you are not putting any risk on your overall projects and teams and resources. Yeah. An all or nothing approach would have required a lot of internal support.

And me and Glenn have talked about this quite a lot, and we're quite confident that would have failed. So our phased rollout is something we would really recommend for everyone going forward. So I think that's it. Thank you very much for your time.

GLENN LLOYD: Thanks, everyone.