Engineers spend too much time fixing broken data instead of focusing on high-level design. This inefficiency, known as workflow friction, is a major enemy of modern construction projects.
This episode explores how intelligent automation and verified software can eliminate manual rework and enhance global collaboration. We discuss the crucial challenge of trust between engineers and fabricators, the necessity of specialised AI solutions, and the innovations of Trimble.
Our guest is David Sanderson, a former site engineer who now leads Product Strategy for Trimble’s structural tools. He offers a unique 40-year perspective, having witnessed the digital transformation of engineering since 1982.
THINGS WE SPOKE ABOUT
● Accidentally invented structural software in 1982
● Lack of trust creates workflow friction
● Specialised AI is best for tedious tasks
● Commercial value prioritises software features
● Contractual deliverable is still 2D drawing
GUEST DETAILS
David Sanderson is the Senior Product Manager for Structural Design Products at Trimble. His expertise is rooted in structural design, leading product strategy for major tools like Tekla Structural Designer, and championing the use of verified, quality-assured software in the industry.
Connect with David on LinkedIn
QUOTES
● "The contractual deliverable is still a 2D drawing, believe it or not" - David Sanderson.
● "The biggest issue that I can see now is that within a project, you've got many stakeholders" - David Sanderson.
● "People have got to start accepting other people's models on trust, in my view" - David Sanderson.
TRANSCRIPTION
For your convenience, here is an AI transcription:
Dusty Rhodes 0:03
Right now, on AMPLIFIED.
David Sanderson 0:04
Will AI ever replace the role of the structural engineer? I don't believe so, and I'm not convinced that it will - we still need structural engineers to do what they do best.
Dusty Rhodes 0:24
Hi there. My name is Dusty Rhodes, and you're welcome to AMPLIFIED, the Engineers Ireland podcast. Sometimes in engineering, you can find yourself spending less time designing and more time fixing broken data. In a perfect world, software should handle the boring, manual task, automatically, leaving you free to focus on high level engineering. Our guest today is an expert in how Intelligent Automation can fix the workflow. He is a former site engineer who now leads product strategy for Trimble's structural tools. It's a pleasure to welcome Trimble's Senior Product Manager, David Sanderson, Hi, David.
David Sanderson 0:58
Hi.
Dusty Rhodes 0:58
So, listen, tell me, David, just to kind of give us a very brief background of why you got into engineering in the first place.
David Sanderson 1:06
Okay, so I graduated in as far back as 1982, when there really wasn't even any computers, true. So I ended up working for a consultancy in London, and I was working on a scheme for a flood alleviation in Ireland on the river Blackwater, and there was this computer in this small room, and I thought, well, we maybe can use this on this job that we're doing now, because We're having to calculate these backwater analysis every Monday, because the Department of Agriculture in Northern Ireland kept changing the flood queue every week, so we had to do this recalculation. And I asked the partner, I said, Can I use computer? He said, No, it's just a white elephant. We could keep that in there, just to show clients we're sort of in the sort of computer agent in 82 so basically, I took it home, wrote some software, brought it in on the Monday morning, when the senior partner brought in the new flood queue, and there was two other graduates that were working with me, and it was a week's work to calculate the main river and all its tributaries. And I said, "stop what you're doing. I've written a piece of software", and we basically were able to compute the whole scheme in 30 minutes, even back in 1982 so went in, said, we've done it for this week. He said, You can't have done I said, Yes, I wrote this computer software. And he basically said, what else can you do?
Dusty Rhodes 2:50
What was the reaction of people? Because they kind of, our computer is just for show, we don't use it. And then you actually gave them a real, practical world use. What was their reaction?
David Sanderson 2:59
Well, they gave me a pay rise straight.
Dusty Rhodes 3:03
Bravo. That's unusual, yeah.
David Sanderson 3:06
And basically they said, What else can you do? And some of the other regional offices started making requests. So I went off out to Ipswich and did a road alignment programme down a wadi in another office, I wrote some 2D analysis software and some flow net seepage through a beat. We're going to put an interceptor culvert. So we had to use the flow net to decide the best level to pull the interceptor culvert at the right location and size to take all the water, because on the other side of the beach, the local town was flooding, so it was just all little practical things like that. And then another office said, Can you write some plastic portal frame design software? And I thought that's getting a bit complicated, even for me. Now, in the early 80s, there used to be these big computing fair exhibitions at the Barbican, where the first sort of BIM and modelling things were starting to appear. And I found a company who had their own plastic portal design software. So I thought, This is ideal. Went back to the office, wrote up to the partner, said, We need to invest in this piece of software. He goes, How much is it? And that was the only bit of information that I didn't get. So I burned up the company. I said, How much does his software cost? We're interested in buying it. And he said, the director, or the Sales Director at that point, said, You sound as though you're from up in this part of the world. Do you want a job? I got offered a job with them, and I went to work for this company called CSC. Well, the first job I ended up doing was writing some portal software. Yeah, because by that time, the first sort of MS DOS was starting to happen, and the sort of standardised operating system for computers. So there was a new design code for steel coming out into the UK and Ireland called BS 5950, so I got a contract from the BSCA to write some software, because engineers were used to just working with hand calculations and things like that. So the idea was that it would help engineers take the software into use if they had some sort of software to help them and understand it. So we wrote some software for simple beams, for simple columns, for composite beams. We also wrote the first sort of 3d building modelling software at that time. And the other thing was this portal frame design software, because it's portal frames use on low rise industrial buildings for all the industrial warehouses and buildings like that. So they wanted a version of the software from the sort of old HP 9816 which they developed it on to re modernise it to a DOS application.
Dusty Rhodes 6:17
Can I ask, David, did you get into engineering as a computer engineer?
David Sanderson 6:21
No, I did my degree in structural engineering. I absolutely hated working with computers at college, so I fell into it by accident. I could see it was a way that I could contribute to the industry. Because if I worked and just stayed in an engineering company. I'd have worked on like half a dozen projects in my lifetime. But through computers and through using computer design, I'm able to influence our companies, our the teams that were developing it wasn't just me. We're able to influence many projects and provide us good tools for engineers.
Dusty Rhodes 7:02
So if it was structural engineering that you got into first, and you say very flippantly, modestly, oh, I just wrote some computer code. You know, you don't just pick up and write computer code. I mean, was this a hobby that you had on the side or..?
David Sanderson 7:18
Yeah, well, the first bit of software was for the backwater analysis was hydraulic software. I was just a dabbler, really. I had a microcomputer that I invested in myself at home, and I just taught myself to programme basic code.
Dusty Rhodes 7:37
You were a structural engineer, and you kind of just played around with writing computer code. Now, the reason I'm saying that is because this may have been quite some time ago, but history is repeating itself right now, where we have AI and the likes of, and I'm thinking of Claude AI particularly, or Gemini, where you can actually speak to the AI, literally, if you've got dictation, and you can say, you know, I want to write a programme, but it's going to do this, this and this and this, and then it'll come back and it'll say, well, here you go. And they go, Oh, that's not bad, but you're missing out on such and so. And you can, literally, I think they call it vibe coding or something now, whatever, but anybody can do it. So if you have an idea, or if you have a problem, and you kind of think, I think I can solve this. Have you played with this kind of stuff yourself?
David Sanderson 8:22
No, I haven't. I have seen it, and a colleague of mine has demonstrated and we were able to very quickly calculate the deflection of a simple beam with a uniform load. Yeah, and the answers were reasonable, yeah. I've not played it myself for two reasons. Is that one, I'm a product manager now, so I don't get my hands dirty, I've crossed over from the dark side. You're in management now. Is that what you're trying to tell us? Yeah, I drive the direction and backlog of the product. Yeah. Can engineers use yes, they can, they can do this. Will it get better? Yes, it will get better and more powerful. But what is that replacing? At the today's it's replacing spreadsheets, where a lot of engineers work in spreadsheets or hand calculations, but working for a company like Trimble, what we offer over and above. That is that we give the customer verified software. We're always updating the calculations, because the design codes of practice change like we've got second generation Euro codes coming out. So that's quite a big lift for us to add in the second generation Eurocodes to our products, and it's a lengthy process to specify, implement and test to deliver quality, assured software with a high level of documentation that the engineer can check that he's got things right. And save in his digital archive.
Dusty Rhodes 10:03
Absolutely. The point I think I was making, more than anything else, is that you shouldn't be afraid to play with something and just see where it leads. And yeah, I think for a lot of people listening, they may play with I mean, you're never going to get it to the standard that you do it using just an AI, but it might give you the ideas. And actually, when you have those ideas, then you start searching around the world to see if a solution already exists, and that's where people find solutions. That side of things, but it's more your curiosity and your willingness to try and solve a problem led you down an unexpected Avenue with computing, and that ended up becoming your career. Now you said that computers drove you nuts and that you weren't crazy about them. What was the exact moment in your life? Where were you and what were you thinking when you said, Okay, that's it. I'm just gonna go build software.
David Sanderson 10:51
I think it was when I joined CSC, who were later acquired by Trimble, that I could see the commercial side of things, and that was of interest to me, and I could then see that I wasn't just working within one company, providing solutions or it's like writing a spreadsheet and then sharing amongst all the employees within your company. I saw that there was a way that many companies were using the tools that we were providing and and that was when sort of light bulbs went on, and how we were making engineers lives easier, and how they could actually save time and concentrate on more in delivering better engineered solutions, and also they could look at more alternatives more quickly and offer better solutions for their clients.
Dusty Rhodes 11:48
Let's talk about workflow friction, because I think you're kind of alluding to it there and reworking it. Oh my god, it's one of the words I hate most. Do it again. You know, kind of, I think we're getting an awful lot of this because problems aren't being solved, because software is not connected, a lot of disconnected software, and things will be done over and over again. Do you think this is one of kind of engineers biggest enemies at the moment?
David Sanderson 12:12
Yeah, that's a really good question. And insight there. The biggest issue that I can see now is that within a project, you've got many stakeholders. You've got the architect, the engineer, the fabricator and detailers, and what you don't see is information sharing, like the fabricator won't trust the engineer's model, so they will build their own model. The engineer is looked at to share information with the fabricator. And you know, if the fabricator is asking for connection forces, they'll take them out of their structural model, and then they will add 1015, 20% onto those values. Well, just in case we've got something wrong, we make everything 50% more than it actually is. And don't forget, when you build your models and you put in all your factors of safety and material on your loads, we've got factors you've also got factors of safety there. So what ends up happening? Information gets passed to the fabricator in this case, and he's working with those loads, so we don't get very well engineered solutions, because everybody is sort of scared of making a mistake.
Dusty Rhodes 13:35
Well, there is that, yes, there's so many different kind of similar data sets, but aren't actually the same, like if everyone's working, how do you solve the problem?
David Sanderson 13:45
I don't thinkI don't think it's ever been solved to the satisfaction that I would hope it could get to. Things have just got to change in the industry. There's got to be some trust. People have got to start accepting other people's models models on trust, in my view, and we are starting to see some of this in the US with our customers, that an engineer is now starting to share models with a fabricator, and he's benefiting from that, because he can roughly, sort of see what materials he's going To need to order in and having a much earlier, even before he's done all the detailing.
Dusty Rhodes 14:27
Yeah, yeah, I get that.
David Sanderson 14:29
A fabricator is suspicious of taking an engineer's model, because if there's a dimension that's wrong and everything gets fabricated to that wrong dimension, then it's a very expensive mistake to fix, you know, is it solved by having more sort of D and B kind of companies that are responsible for the design and build. But then what do all the engineers do? And, yeah, I'm just thinking steel here. You know, we've got the world of concrete as well. You know, we've got engineers designing. Concrete Structures, fabricator, are they? Contractor? Then constructs solely on site, and probably does all the detailing as well.
Dusty Rhodes 15:10
So then, when you look at this kind of on a global scale, where you've got one party in India, another party in Australia, another party in Dublin, all working on the same product, can I ask you about trim BIM, which is something that you guys are working on. Does this kind of help reduce the latency in global project collaboration?
David Sanderson 15:28
Trim BIM is a Trimble invention. Basically, it's like souped up IFC. IFC is international industry foundation classes for sharing information. It can be quite slow when you move in large models between products. Our technology team within the techler organisation came up with this souped up format, which is souped up IFC, and we're using it now for our workflows for transferring information between products. It's just a vehicle for sharing information. At the moment, it works within the Trimble domain. The Trimble products that I sort of interface with. We've got a trim bin connection between these products, okay?
Dusty Rhodes 16:20
And does that help? Then what we were talking about, kind of about people not wanting to work off other people's measurements, or that kind of trust factor.
David Sanderson 16:26
We need to transfer information reliably and accurately between products. And I think once companies see that happening, they're more likely to accept, say, well, I won't rebuild the model. I'll take that model into use. Does it reduce rework? Hard to say, but what it does do thinking about steel. So the engineer designs his model is sends it off to the fabricator. The fabricator then starts redesigning and making changes. He then needs to send that model back to the engineer, who then verifies and said, yeah, those changes that you've made are going to work. You know, it just makes that sort of whole cycle a lot smoother. As well as passing the model, the fabricator and the engineer can attach notes to changes. And then when the engineer opens it up in his model, he can see change made by such and such on such a day.
Dusty Rhodes 17:33
And then it just makes it all a lot easier to handle.
David Sanderson 17:34
Alright, if you don't have these sort of tools, then it's like a phone call or it's an email. And then how do you keep track of all the phone calls and the emails and get them in the right chronological order the changes?
Dusty Rhodes 17:48
So we have exactly the same problem, and we just make podcasts same thing, like, you know, how do you keep all this information, all these different people and just have it in and we have one project management software that we use, but not everybody puts their comments or their notes into it, so when other people are referring to it, not all the information is there. So, I mean, I completely get anyway, let me go on a different track. Listen, we're talking about computers and all these things that can help with engineering. We can't not talk about AI, okay, because it's just everywhere, and people are starting to use it more and more, and we're getting more familiar with it. How do you ensure that AI complements human judgement rather than replacing it?
David Sanderson 18:27
Right? This is the $64,000 question that I'm getting asked virtually every day of my working life at the moment. So you've got to look where AI can be best used. And at the moment, AI is best used for dealing with repetitive, tedious tasks. So will AI ever replace the role of the structural engineer? I don't believe so, and I'm not convinced that it will do. We still need structural engineers to do what they do best, design the structure, understand how the structure is working, and I don't believe AI will be able to do that, but where AI could be useful is helping the engineer find his way around about the tools that he's using. It might be an assistant within the software. It might be some sort of repository or some sort of large language model where all the project documents have been put into this large language model so that the engineer can say, oh, there was some change by the architect on floor four. He's changed the whole flooring system so he could go in and ask, what are the new loads on floor four? And rather than searching 1000 project documents and or looking in a binder or some folders, he can use the tool for getting the latest project information. Another. Area where it could be used is, once it moves sort of over to the fabrication side, you need to generate a lot of project, documentation drawings, fabrication drawings, and there's a lot of work going into even though the drawings are automatically created by the system, they all need sort of cleaning up and tidying up. So companies have standards. So you could envisage a sort of fabrication drawing AI to go through all those drawings or use it to base other drawings on, very efficiently. So the tools like that, tedious things where you spend a lot of time on the project deliverable, especially around drawing basically, the temptation is very easy for somebody who's using, like an online AI, like, you know, Gemini or Claude or something like that.
Dusty Rhodes 20:44
Do you think it's a good idea for people to kind of try and come up with these solutions, as you just outlined there, by using kind of a bog standard bugs? Am I already saying bog standard, or the miracle that is AI sometimes to use something kind of generic, if you want to put it that way. Or should we be developing ais that are specific for this, for the different strains of engineering?
David Sanderson 21:16
I think the latter, because an AI is only as good as the data that it's got to work with, and the more data you have, the better outcome you can get out of AI and I know you need specialist ais that focus on tasks rather than something generic, because you're not going to get a very good outcome out of a generic Yeah. So it's more data, more specialised AI solutions. Yeah, that all can be built up from some base AI tool. But then if you're doing something specialist around structural engineering, then you need to, as I say, provide specialist data and train it in a way that's suitable for structural engineering.
Dusty Rhodes 22:00
Speaking of designing and training, I mean, you're working with software all the time, and I'm sure you get millions of suggestions. Why doesn't it do this? And why doesn't it do that? Can we have this feature and that feature and all that kind of stuff, you know, kind of you're leading the pack there. So how do you decide which feature requests get prioritised and put into the roadmap.
David Sanderson 22:24
Well, obviously we have a collection system in our company called JIRA. So every time a request goes in, it gets logged in JIRA, and if the people log in it are really good, they will search JIRA first and add another request or a similar request from a different company, so we can list the companies that have asked around. This feature. There are some AI tools now which will search the JIRA database and look for commonalities. And then we have certain initiatives. And as my role as product manager is to sort of look at where the direction of industry is of what people are doing so and then we try and as we open up bits of code or we're adding new things, we'll look and take things from the backlog and add them in. But basically, we've got to do a lot of sums, basically, and say, Well, if we had this feature, and it seemed at either the UK or North America, and there are so many engineers there, and we've got so many existing users, we've got to come up with a sort of Euro value of What we think the software is going to generate by adding, adding that initiative or that of that feature, and it is really hard to do, but there's some things, like the second gen Euro codes that we've got no choice. It's an essential part of the software. We've just got to deliver that and have it ready for when it becomes mandatory.
Dusty Rhodes 24:03
Let me ask you about one of the innovations that's out in the world at the moment, and that's the integrated portal frame design. Some people have called this a game changer for industrial sheds.
David Sanderson 24:15
My history with portal frames has gone right back to the sort of mid 80s? Yeah, we've had three or four generations of portal frame software, and it was in a separate tool, which sat beside the main Tekla structural designer tool. And then you've got all the problems with integration. So you could model portal frames in Tekla structural designer, but then you had to take them out to another piece of software to design. So one of the things we've done with our 26 releases, we brought in the design solution into the main body of the software, so that we cut down on this necessity to have this integration. So. Why are portal frames important? Because 50% of the UK fabricated steel goes into portal frame buildings. A lot of our customers are designing portal frames. We've got a very efficient elastic plastic solution that designs these and even though it's a 2d design problem. Working with between 2d and 3d we can offer our customers a very efficient design tool for these buildings. And you know, just you go on to these shopping parks or these out of town centres that it's just full of sheds that people wander around in and buy stuff. They're all over the country, travel up and down the motorways. All these big distribution centres are all portal sheds, basically. And the other thing was, I've got all this inherent knowledge about portal frame design, and I'm close to the end of my career, so I needed to pass all the information on to my colleagues.
Dusty Rhodes 26:05
Let me ask you about that. All right, because, I mean, you got into the business before computers were really a big thing, all right. And literally, by accident, you fell into the future. That's the best way I can you saw this computer. Oh, hang on a minute. Boom, and you just dived into it, all right? And since then, I mean, the world has transformed since then. I mean, you thought that was magic. Then what you could do with the computer and the problems it solved, then the internet came along. Oh, my God, jaw drop. I've just sent a message to somebody. I remember the first time I did it, all right, I've sent a message to somebody in Vancouver 10 minutes ago, and guess what? My wife's gone what I said, they've just replied, You can't do this with a post. So the internet kind of was just like, wow. And then AI came along, whatever, two, three years ago. And again, jaw dropping. What do you think with all of these changes and everything that you've done over your career, do you hope that your lasting contribution will be to the structural design community.
David Sanderson 27:05
Just that I've helped my fellow engineers, just made their lives easier in the sort of 4040, plus years that I've been involved, what we've not been able to achieve, which is a disappointment, is that the contractual deliverable is still a 2d drawing, believe it or not, and we fill all these wonderful 3d models, all These wonderful BIM models, with all this information we've discussed the sharing, maybe, maybe not. But you would think that this day and age that we would be able to deliver our companies who fail to deliver a 3d model and say, build from that ref. But no, that's just an adjunct. There's still the contractual deliverable is a big pile of 2d drawings. And, you know, I remember being on site opening big, sort of a naught drawings, trying to fit rebar into some poor and the rain wind blowing and blowing it out, and things were getting dirty, and I couldn't read the text. I know some very advanced companies now are using like iPads to put the drawings on and things. So you can carry an iPad round. But even though we've done all this wonderful stuff and this computer aided design and developed these clever solvers that will solve big problems. It's the fact that we've never changed this contractual deliverable of a drawing, and that's like in 40 years. So hopefully in another 1050, years, we might have got somewhere with that.
Dusty Rhodes 28:59
Do you know what that is? An absolute, fascinating point that I want to finish on, because it's just something that I think is worth thinking about for the remainder of the day. After everything we've been through, the work is still delivered on paper. Wow. If you want to find out more about David's work and the latest 2026 releases and innovations, you can check out the links for Trimble and the product suite in the show notes below on the podcast player right now, but for now. David Sanderson, thank you so much for joining us. You're welcome.
David Sanderson 29:29
It's been enjoyable. Thank you.
Dusty Rhodes 29:30
We hope you enjoyed our conversation today. If you know another engineer who's interested in the future of AI and structural design and all kinds of things engineering, please do share this podcast with them. They can find us simply by searching for Engineers Ireland, wherever they listen to podcasts. This episode is produced by dustpod.io for Engineers Ireland, for advanced episodes, more information on the latest trends in structural software or career development advice. You can find a wealth of resources on our website at engineersireland.ie, until next time for myself. Dusty Rhodes. Thank you for listening.