Can AI Take the Repetitive Work Out of Quantity Surveying?

Nina Young: 00:08 Hello and welcome. You’re listening to This Is Surveying, the podcast shining a light on the people, ideas, and stories shaping this incredible profession. I’m Nina Young, founder of Surveyor UK and the Surveying Room, the community bringing surveyors together, breaking down silos, and making surveying visible. So for now, let’s dive into our latest episode. Hello everybody, and welcome to This Is Surveying. I’m pleased to welcome Will Dole today. Will is a former quantity surveyor and has spent the last eight years as the founder of Gather, which is a leading platform for quantity surveyors in the UK. I also have followed Will quite closely, especially over the past year. He has an exceptionally good uh LinkedIn newsletter called The AIQS, which has become very popular. And it’s also something I recommend people uh take a read of. I’m gonna give Will the chance to talk a little bit more about himself, his background, and we’re gonna dive into a lot about AI today, which is uh obviously a topic most of my interests and Will obviously as well. But without further ado, welcome, Will. Glad to have you on here today. Will Doyle: 01:26 Thanks, Nia. Thank you for the introduction. Yeah, looking forward to kind of discussing everything about kind of quantity surveying and maybe touching a little bit on AI. I don’t pretend to be uh to pretend to be much of a of a an expert in how line songwriters models work, but I can definitely see the applications for for quantity surveying and and the wider sector already. So looking forward to discussing that a little bit more. Nina Young: 01:48 Super. Nina Young: 01:48 So before we sort of go into more around sort of gather and and and what you’re doing and have been doing for the last number of years, with regards to being a QS, did you fall into QS or was it something you just thought, I want to do it? Did you wake up one day, want to be a QS? Will Doyle: 02:05 Um I did this is like my um university application letter, actually, this is a story. So I I grew up in Manchester, um, and I grew up when Manchester was being rebuilt after ’96, and what happened right in the centre of town, and then grew up in construction ready. I I I’d love to come into Manchester to see it change, and it still is changing all the time, 30 years later. So um I kind of uh applied to to work in Quite Surveying because I was always good at maths, kind of liked construction. I thought that those two kind of aligned, but I actually fell into a sector that I didn’t really predict that I would do. So I I when I went to Loughborough, when I did my point of surveying degree, um, you have to be sponsored. So I I wrote to lots of different construction companies and I’d seen Valphabiti everywhere. And it’s like, yeah, I’d love to work with Valphabiti and do lots of work in the centre of town in Manchester, really kind of exciting place to be. And I I don’t know what I did, but I I managed to send a letter to um their office in Derby, which was in in rail, which I had nothing, no interest in or nothing to do with what I actually thought I I liked and has given me the best career. Worked all around the world, across the UK, a little bit in Hong Kong, Malaysia, a little bit in Singapore, and just alphabetic was one of the best jobs I had, and I did it by accident. I I sent their letters to the wrong place and um that’s interesting. Nina Young: 03:29 So wrong place it went to. I mean, they’re such a well-known company as well. I mean most people know who they are. Did um did uh one of the things I perceptions, or we have a lot of people listening that won’t be quantity surveys, for example, and we have a lot of people that are thinking about doing surveying with quantity Nina Young: 03:50 surveying. Now, there is this of the the people have certain perceptions or stereotypes of what it is. Is it predominantly a couple of things? Is it desk-based? Is a lot of it also dealing with people, or is it very much head-down data analytics forecast, you know, that kind of thing? Will Doyle: 04:08 I think it depends the sector, depends on on the experience. And I think it’s changing. And my my first experience with being a quantity surveyor was brand new suit, first job on the uh for Balfaviti, Thames Link in London. I spent the first week sat on the floor filing paperwork and going through shift records and scraps of paper and putting them into the right folders and things like that. So that was interesting. And then I spent a lot of time out on site understanding what we’re actually building, which is I always recommend to someone who’s who’s new into any role or any any sector. And I think the role of quality survey is just kind of gone from being very site-based and lots of people, uh, interactions and face-to-face and meetings and all the kind of soft skills that are really important there, which we’re traditionally not very good at in construction, but we’ll we’ll move on from that. And then we’ve gone to this kind of like almost like we we’re glued to our desk. We’re doing spreadsheets, we’re doing application payments, we’re writing letters, we’re looking at cost to complete forecasts. And it’s all been very admin-heavy. And basically, you have to be good at Excel to be good at QS. And I think that’s always kind of needed to need to know your numbers, but we’ve all been like XL, XL, XL, and whatever tool you use. And I think that’s actually going back the other way due to kind of using AI and other kind of more automated processes now. We’re actually going to be spending less time at our desks, less time doing spreadsheets and more time getting back to talking to each other. So I think that’s a really good thing. So I think QS is a very different and it is changing. And I think if you get into quality surveying now, probably you there’s two ways to do it. You need to be really good at data, and then you need to be good at understanding how they different bits of data fit together and and using that data. And or you need to be really good at talking to people and really good at soft skills. I think those are two things that you need to be a good QS. Some of the best QS’s that I’ve ever met haven’t come from nutritional backgrounds. I know one lady who is probably the best QS that I’ve ever worked with trains a hairdresser, but she knew exactly how what what the job was, how to speak to people. Nina Young: 06:15 How to speak to people. Yeah, that communication skill is one of the ones that gets mentioned a lot on the podcast as surveyors say it’s it’s kind of you’ve got your technical side, but then the actual soft skill scot side is just as important. You’ve got to convey either information in a way people can understand, clients understand, but also sometimes deliver difficult uh you know, have difficult conversations and difficult things. Now that’s interesting. And Nina Young: 06:39 I think with regards to sort of standing back in generally what you’re seeing across across the you know construction QS with regards to the AI influence, how have you seen things sort of play out sort of the last year or two? What are you kind of seeing? Is it increasing? Is it challenges, risks, you know, kind of what’s your feel for what you’re seeing? Will Doyle: 07:01 There’s a lot of kind of negative, there’s a lot of nervousness. I presented a seeker event, so then Jake Contracts Association event on Wednesday. And there’s a lot of people very nervous. A lot of people talked about AI slot, which is a great phrase. A lot of people were just getting pages and pages of letters that don’t make sense and not written correctly, got contradictions and huge presentations or just like overwhelmed by just AI outputs, and usually from kind of more junior people, because they perhaps not know any different. They can’t really check what they’ve said because I’ve got to do it the hard way. Because then I’m just, yep, kill. And the likes of Chat GPT in this world will then fire out me at a letter. Yeah, that must be right, because that large language models are in it. And I think there’s a lot of negative stories there, but it’s also a lot of positive where we see a controlled use of tools and software and large language models to do things a lot quicker. I think if you’re in the sweet spot of someone who’s probably five to 15 years experienced, and you’ve probably done the hard yards, know exactly what you you should be saying, and just found an issue where you have to work 60, 70 hours a week to kind of keep up. Well, this is this is amazing, isn’t it? Now I can do my 40 hours as in my actual contract and do my job as really, really well, and probably do it more efficiently because I’m not going to do the number crunching the things that used to take me lots and lots of time. So a huge benefit there. But I do worry about people that are new to the sector because they’re not they haven’t got the opportunity to do the hard yards. And I I we probably had this in a different sector, probably 20 years ago. And one of the best engineers I’ve ever worked with, he was very passionate that people shouldn’t just go straight to AutoCAD when they were uh a graduate engineer. His view was um you can have to get your laptop out until actually this calculation needs to change. I can do in my head and work out, yeah, that’s about right, because I can do it from first principles, and that’s where we’re gonna perhaps get to with the the next generation of surveyors, where we just rely on software, we just rely on that, and we don’t have the the the theory or the fundamental foundations to understand if that’s right, and that’s a little bit of a concern. And that came up a lot. Nina Young: 09:08 Yeah, I bet Nina Young: 09:09 it did. And I think I I I read one of your actual um your articles about that as well. I remember you touched on something around that because and and something I wrote about recently was also around the whole the judgment paradox, which was where you know, I think judgment is going to be absolutely the one of the most crucial skills ever in all professionals in work, I think long term with AI. And it’s like, how do you develop judgment when you don’t do all the entry-level stuff you don’t learn from your mistakes? And then the entry-level roles are being replaced with you know automation as AI software, etc. etc. How are they gonna learn? And this is playing out already in legal, uh, paralegals and trainees. How much of the and it’s like, and and so many people are now talking about that. It’s like, and everyone’s coming out of university now, they’ve used it for three years, they expect to use it, and then they’re worried about getting a job. But then, like you say, it is that sweet spot, and this is the the irony, I think, of it all, is those surveyors that are a bit older, wiser, they’ve got that judgment experience. There’s a lot of those that are also resistant to AI, whereas actually they’re the ones best placed to question the outputs. Will Doyle: 10:15 Exactly. That’s the kind of paradox you described. I think it’s quite exciting, but a little bit scary at the same time. I think it’s quite exciting. If I was a practitioning quite surveyor now, I think one of my friends says you’re quite dangerous because you you kind of know what you’re doing, but you can also now do it a hundred times faster than you you did before. So there’s obviously the opportunity to make mistakes is is is greater there. But you can now be more efficient, you can do all the things that you you wanted to do, but just didn’t have time or that. So we’ve been really pushing our AIQS and and I’m not going to push the product, but we’ve talked a lot about the time savings and the and the how it’s cheaper than a human equivalent, which is an interesting angle. But what a commercial manager said to me last week, commercial writer, sorry, is that those are not the benefits, Will. The benefits is actually we can deliver our commercial handbook, we can do what we said we can do, because we’ve never had enough resource, never had enough time in the day to do these hundred things that we’ve got that said that our commercial processes, we just never get around to them. And now that’s where we’ve actually had to do a more thorough job because we remove the constraints, the time and the data entry and all that sort of stuff. So he views AI as a way to him for him just to do what he should have been doing. Nina Young: 11:26 You touched on um yeah, that’s a good point. Because you you touched on AI slop and a well as well. But something that I’m hearing increasingly very recently is around, you know, surveyors are saying, well, actually, using AI, we actually find is more accurate than a person. And it’s I I think if you use it properly, and and we’re seeing this across other professions, medicine, everything, because the way that AI can capture the data, analyze it at volume and scale, and it doesn’t get tired and all this kind of thing, doesn’t get hungry, doesn’t get distracted. I mean, you think about the taxis like Uber taxis and you know, the way that you know driverless cars have been now proven to be safer than actual taxis with people in them. And I wonder if that you’ve seen that or that’s uh something that you’ve seen in your your own experiences. Uh because it’s quite it sounds really controversial because everything seems to be AI versus humans when it you know AI is like has been developed and created by humans. It’s like it’s it’s humans’ responsibility at the end of the day. Will Doyle: 12:27 I drive one of those cars you mentioned, and just statistically, it’s the safest car on the road in Europe. Nina Young: 12:33 Yeah. Will Doyle: 12:33 It would be even safer if everyone drove them. And there’s a there’s a stat around it. People driving that same car manually and and still safe. But when they drive it on autopilot to a degree, not hopefully autopilot in the UK, it’s even safer. And isn’t that the same with any profession? We can like humans are good, but when we have days off, we forget things, we we’re busy, we’ve got other things we’re on our mind, and a large language model can do things in our routine, they’ve got guardrails, they should be better. If we’re if if they’re not, we’re we’re not doing the right thing here, are we? Like, and we’re not talking about the quantity survey as a whole, we’re talking very specific tasks. We’re not saying let’s replace quantity surveyors in full, we’re saying let’s replace the arduous, horrible stuff that when someone wrote their application to go to university, they go, I’m gonna get into quantities of fame because I love going through paperwork. No one says that. Uh, but it is important to those hard yards. And going back to your earlier point, I do think that’s gonna be because I I started this business because I spent years going through site records. I was on the wrong end of a really bad claim where we’d have the right site records. Like I’d been through that pain, and the people that I speak to and sell gather to and love gather and adopt it are the people that have been through that same pain. And I feel like you do have to understand that problem and and feel it. And then when you do that, you kinda you know when you automate something and you you let it run that you know that it’s working because you know the complete opposite and how how negative that can be. So that’s kind of where I see things going. Let’s let’s do that, let’s re replace the horrible, boring stuff that no human wants to do and let’s do it better. Nina Young: 14:08 I think it yeah, I was a good point. And before we we we go on to discuss gather, I think it it does come back to that whole point around the fact that if you use it correctly, it will be more accurate and it can do things quicker, faster than humans can. There’s been a lot of talk about, you know, very much on social media and online and uh speaking to surveyors and firms, you know, AI, there’s a lot of narrative around AI is not gonna replace surveyors. But I, I don’t know, your view, my view would be that we need to kind of move on from that because that’s just kind of protecting the way we’ve always done things. But really, we need to be talking about AI is gonna help remove all that mundane stuff and help free you up to do the high quality, the judgment, the interaction, managing relationships, projects, clients, that you know, interpersonal skills and all that kind of thing. And I think I don’t know what you’ve seen from your side, if you’ve seen a still the very much, like you said, a negative view. Is it because of the AS LOP or is it people’s identities being challenged or they’re worried about their jobs? Will Doyle: 15:12 I think it’s difficult, isn’t Will Doyle: 15:14 it? And people are scared of the unknown. I I’m sat in Manchester, I’m sat in Ancoats and Manchester’s four or five mils, the other side of that wall, where we the heart of the industrial revolution, which is which is Manchester very proud of. And yes, we we but it was going up on steam, and then when someone made electric motors, that have we we put those in in the factory, and it was better, but we eventually had to redesign the factory and to around the power source. It was no longer this big steam thing that sat in the middle. We now can we can put the machines anywhere because we just got electric. Isn’t that the same what was in with kind of AI and according to the main or any other profession? We we’ve we’ve now got our profession we’ve always done, we’ve done really well, and now we we’re using kind of the AI model to make us faster and quicker and more efficient than all the other adjectives you want to use. But we have to get to a point where we need to redesign a factory, and that’s kind of where we’re getting to, and we we see that in software development. Like we’re a software business ultimately, and there’s a lot of money in software businesses, there’s a lot of investment that goes into it. And the first kind of use of large language models was just writing the code a little bit quicker, right? The the the the writing the code and tapping the keys was always the hardest thing. That’s the const the constraints change now. The constraint is do you know what the problem is? Can you think about system architecture better? Can you make sure you can really kind of react and understand the customer and spend more time on site? The constraint is no longer how fast can you type. And how do we apply that to construction? How do we apply it to quality surveying? Well, it’s not about how you’re having the best spreadsheet or all that. Isn’t it about understanding your customer better, understanding what they want and the constraints of your contract and and and the infrastructure you’re building or the project you’re building, and then using data, not just how how how good’s your spreadsheet, or that give me the best cost to complete, or the best forecast. So I do think we’re gonna go through the same sort of stage revolution that we we saw kind of in in Manchester, where we have steam great, well done. And that’s a fantastic change, but we need to redesign the factory. Nina Young: 17:13 That’s that’s that’s a great analogy. Yeah, I like that. I’m probably use that somewhere, Will. Will Doyle: 17:19 I’m not taking full credit to that. There was an economist such I I I I was trying to get the article on my notes, but someone wrote an article about this probably 20 years ago about redesigning that. And when you now apply that, and I’ll send it you so you can perhaps put it into show notes and yeah, please. That was that stood stood with me because we need to get to the point where do you need to just fundamentally change the way we’re working? Nina Young: 17:39 Because at the end of the day, it’s like uh I see that a lot there’s a lot of resistance because it’s th threatening professional identity across all professions. And but we need to focus on the outcome. The outcome at the end of the day for the client and doing something better, quicker, and maybe sometimes, you know, the the talk about costs going down and things like that, I think is a good Nina Young: 17:59 thing. So moving on to to talk about gather, now that’s you’ve been you founded that about eight years ago. But sort of, I think you touched upon kind of wh why it might have come about, but can you sort of explain why you started Gather? Will Doyle: 18:16 Yeah, so I think I kind of grew up in in a career perspective of just understanding the importance of of records. I think records, records, records is a phrase that most quality surveyors will know. But I just felt like that was drummed into me and and when we had good records, we had a good project. And you probably don’t need good records on your good project. That might change a little bit now. And then but ultimately people wanted records when something went wrong. And I seem to, I’m not sure it was me or projects in general, but I seem to be a lot of projects that went wrong. And when we didn’t have the right records, then we the outcome was was pretty poor. Like we ended up in a huge claim on a project very close to where I am today. And going through piles and piles of records, trying to actualise what we did and understand why it cost extra money and try and pin the blame on different organizations. So it wasn’t all us, we’re sure with much humility. And the biggest effect that had on me was that people that I I think trusted me, and I I definitely trusted them. Used to go for lunch with them, used to go to the pub with them. They didn’t want to talk to me afterwards. They didn’t was like, oh you’re like getting into a claim, this inevitable bund fight under every construction project. Like, what what are we doing? And that kind of I left that organization. I went to work for a different organization uh after that project, and the problem was the same. Again, I I do still worry that it was me, but it’s the following you around. Yeah, exactly. But that problem was the same, and and it took me some time to kind of raise the money and the capital to start Gather. And I I kind of registered business probably two, three years before I did any work because I was like, okay, this is actually more expensive than I thought it would be. Um very different if you’re starting now, but back probably seven or eight years ago, we needed quite a bit of capital to start Gather. And was it when I was doing that and I was freelancing to try and slowly work more full-time on Gather, the problem just getting worse. Every single project out there, just a poor records. We’re actually going backwards from an industry that was quite good at writing things down on pen and paper, good at allocation sheets on triplicate pads. And with the rise of kind of bees and people glued to the hand, that we thought it was fine to send a text message or a quick note on WhatsApp. Nina Young: 20:24 Yeah. Will Doyle: 20:25 And the art of actual proper record keeping and telling a story of what happened on a site and using that to explain, okay, this could take a little bit longer, or actually, this was a little bit different to what we thought. Just all lost. And the half the battle when we started Gather was just getting people to understand what they should be writing. Like and we my my my colleague Ben, who who started CMAR, which a lot of people listening will know very successful software. Will Doyle: 20:50 He he for contract manager in NEC, he talks about three things knowledge, culture, and discipline. And that’s how we kind of push Gather, and that’s how we roll Gather out. And we need to have the knowledge, we need to understand why we’re doing things. Why is it important to keep records? Why, why are records why we say records, records, records? Why do you say it three times? Is that important? And what should we what’s the knowledge and we’d have that? What how’s it linked to our contract? How’s it to what we said we’re going to do, our scope? The culture of it is is the hardest thing in construction. We’ve not cracked this and that we’ll always need to keep working on it. It’s it’s usually the why. Why is it important to me? Why is it important to us giving positive feedback? I think when I interviewed lots and lots of people when I started this business, I realized I used to be a horrible coin sway. I was a horrible QS because I only ever rang or spoke to my site manager when something gone wrong. Nina Young: 21:39 Okay. Will Doyle: 21:40 I and where we talk about now, fast forward seven, eight years ago, or for seven or eight years from to now, we see lots of kind of positive thank you. That was really good. That those records helped us do this, so that helped me understand thanks to those folks. And we’ve kind of built that in together because we realized how bad it was when I interviewed people. I was like, yeah, I just filled that record and goes as a black hole and hurts. It wrong, someone will ring me next week and say, Where on earth is this? And so that’s the culture bit. And when we need to get better at that, and and the process is just the software. It needs to be efficient, it needs to be slick. But we need to think about knowledge, culture, and and and discipline. It’s the process, the discipline. Nina Young: 22:18 I guess it’s like anything, anything, any software. It’s it’s you could you can build the most amazing solution, but unless you get people on board to understand why and what how to use it and why it’s important, there’s so many times that even before you get to the software, it’s changing a total mindset of something. And I guess the software is just the discipline. Will Doyle: 22:36 Yeah, the yeah, the discipline’s easy, that’s just a routine. We used to do it. It doesn’t matter if it’s on a spreadsheet, fancy app or whatever. That’s kind of the easy bit. And software environments isn’t easy, but that that once you get to that point, you’ve got a product, that is kind of the easy bit. It’s the knowledge and the culture that is usually the the more challenge, challenging element. And knowledge is okay. We can we can sit down and talk together about the theory about it, and we can we do that all the time. And we people learn and people understand it, and they they like to have the background of why this thing is important. It’s the culture bit that’s really hard. Nina Young: 23:08 Is there a lot of scepticism? Is there what the what is the sort of the main things in the culture that you’re still seeing even now that are kind of you know that we’re just less carrot more stick? Will Doyle: 23:19 I think we don’t like saying thank you goes a long way, and I’ve realized this over the last few years. And and people want to, I was very much just someone’s job. Of course, they’re gonna do it. What we’re paying them to do this is in the job description, but actually explain the why and saying thank you, that’s what really helps me out in my role. Like, we’re in a position where site records, capture site records, is done by someone in the poor range at three o’clock in the morning if you work on a live infrastructure project, and they don’t get much benefit out of it. It’s not like an accounting software where the accountant goes, Oh, this is really good, it’s held me loads of time, and I get the benefit out of it. We’re asking someone who’s very detached from the quantity survey and all the project in my job coaching to do a quite a bit of the legwork, which we do make easier. But we’re asking them to do something that we’re gonna rely on and saying thank you, giving them feedback where yeah, that record you did last week and those photos explain that issue. That’s really helped me do this, this, and this. That’s the bit that we’ve that we people find hard because isn’t that their job already? Why they just do it anyway, but no, they need to know they need to complete that feedback loop. Think about your personal life, you want to know it what the effect of somebody’s done, is it’s just like saying thank you, isn’t it? We we would do that thirsty. So can we do that professionally? Yeah, and that’s gonna learn. Nina Young: 24:27 Yeah, that’s that’s that’s really interesting. And it again it’s coming back to those people’s skills and understanding because it’s that reinforcement, and it doesn’t have to be negative, like you say, constantly you haven’t done this, and like you only pick up the cult when something’s gone wrong. But it’s literally that acknowledgement because I think you know, acknowledging someone, the validation of it, understanding they understand the importance of the work, and if they do it well, what are the knock-on effects? They’re gonna do it more consistently better. They’re gonna they’re gonna want to. Whereas otherwise, it’s very soul destroying, isn’t it? And the only time you interact with the the output of your work is you’ve not done that right. It’s it’s awful. Will Doyle: 25:09 But quantity surveying as a as a function has traditionally been quite a soul-destroying, bickering, argumentative profession. Like we we all right, we’ve seen as these people like kind of knock some money off that, red pen that, there to kind of argue your case and argue black is white. And it’s not really about that, is it? Really? It’s a it’s probably the world’s change. We’ve got a contract format that is meant to be mutual trust and collaboration in terms of NEC, in which most big projects now run on. We’re meant to be more collaborative, which isn’t not a word that many people understand, and I don’t think I fully understand it in practice, but it’s it’s about trying to do the right thing. What’s the the right outcome here? And what I learned from one of my customers when I was doing that claim is we would we quite like working you could because you showed a bit of humility because you would say when something was wrong and sorry, that was ours, that was our issue. Because it rather than me saying, Yeah, it’s 100% your issue, I won’t pay for everything. It made the bit when I said, Oh, these things here, that was our issue, but those other things I won’t pay for. It made that conversation a little bit easier. So I I think, yeah, with a little bit more humility, a little bit more focused on people’s skills. We’re not here to, we’re not professional bickerers and not arguers and give some feedback. I think that’s the way that quantity surveying needs to go. And I wasn’t this, by the way, I’m very aware on my kind of soapbox here saying how things should change. I think I was probably the the inverse of all the things that I just said to you. But I think looking, I got I’m really privileged that I get to speak to hundreds of quantity surveyors every month, different projects, different sectors, different people that like and when I when I see the good ones, the ones that really enjoy what they do and are really effective, they’re doing all the positive behavior, though all the soft skills are good. They’re not just sat in their bedroom sending out angry emails or working on spreadsheets. So and that used to be me. So um, yeah, that’s where I think that it’s changing. Nina Young: 26:57 No, that’s really interesting. And when you think about the way things are evolving with AI and people saying, you know, the human skills and it’s like that’s gonna be that’s even more important then, isn’t it? It’s like when some of the mundane work’s taken away. Nina Young: 27:12 So tell me, um, Gather, you’ve been going what about eight years, like you’ve said. So kind of how many, how many firms do you deal with, quantity surveyors, what’s what’s the skills? Will Doyle: 27:22 So we’re nearly up to a hundred different organizations and they vary from big asset owners and and big tier ones, right down to five or six person specialists and and everyone in between, which is which is quite interesting. Uh like I said, I’m very privileged that I get to speak to lots and lots of different quantity surveyors and industry professionals each month. Um, over 5,000 projects, mainly across infrastructure. So nuclear rail, which is kind of ridiculous where we’ve business started, highways we’re seeing huge growth in kind of the energy sector, utility sector. So the UK. Um, most of our work is in the UK. We do have some projects in in Canada and Australia now, and hopefully some in Southeast Asia very shortly. So we’re starting to see that. Incredible, yeah, yeah. But it is a it’s a challenge because one thing we’ve noticed with Gabriel, and this is a scaling challenge for us as a business is so we see how effective in-person training is for field people. So we’re going back to these soft skills again. We think software sitting in your office in Manchester with your floating egg chairs and you yoga on the roof and all the other weird and wonderful software things that we have. Or do you want to go out on sight and actually speak to people who on the night shift doing this project on middle and in the middle of nowhere? We’ve found in-person training to be so effective around that kind of softer skills, going back to the explain the why, together with that organization and those individuals. And as we scale the business and look further afield from the the UK, that’s going to be a bit of a challenge for us because we know how construction professionals respect that kind of face-to-face. Right. My colleague Tom spent 15 years on site as an engineer, and he gives the best training in Gather because he’s just so relatable, like just so relatable in terms of what he does. And I think that’s that’s really important. So that’s a challenge, but a good one to have. Nina Young: 29:09 No, it is a good one to have. You you’ve done it sounds like you’ve done incredibly well. And I’ve what I’ve picked up on and some of the things that you’ve written about as well, is that obviously Gather wasn’t uh an AI-based platform when you set out, but you’ve started to introduce kind of AI into Nina Young: 29:27 it. Is that my understanding fairly recently? So, what are you using the AI element, sort of how we use that within the software? Will Doyle: 29:34 So Gather’s boring. It’s just we’re actually good records. Let’s just let that. I’m not trying to dress it up because people want to go for drones and robot dogs, all this stuff. I I’d love a robot dog, but yeah, I would. Nina Young: 29:47 I want one as well. Will Doyle: 29:49 And yeah, and it would be probably more easy to do with my current dog. But yeah, but here we go. But no, joking aside, I think gather is once we’re boringly predictable, it’s a good record, structured record, the same record across our entire business every single day. And we’ve got hundreds of houses, but millions actually, of records now from all different organizations. And that our customers own their own data, so we don’t put them all together, it’s a very um, but when you’ve got lots of structured data, which is very objective around how we measure things in accordance with standard method of measurement, and you pair that with anecdotal explanations, this is what happened on site, and this is why we achieved 20, not 30 of XOY. And we also used to build lots of spreadsheets and reports and PDFs and dashboards. But when you’ve got lots of structured data and you’ve got lots of anecdotal data that’s linked to that, large language models are incredibly effective. Incredibly effective at going. Oh, what happened last Tuesday that was different to three months ago when we last visited this site? And just cutting through stuff where you rely on someone’s expertise or download that report. Or Tuesday was the 15th, and we’ll compare that to the 15th of the last time we were there. Oh, I can’t remember when we were last there. Just the kind of that noise around managing a project, and you put that in the hands of people that know the project and know what they need to look at. It is scarily good. And we we we make sure it’s got the right guardrails, it’s it’s kind of founded in in in in good practice. So I I’m a quantity surveyor. My colleague Ben helped draft an EC4. So the the AI offering that we offer, our plug-in that’s available, things like ChatGPT, Claude, Copilot, okay, is is grounded in just good practical knowledge. Like it probably writes a conversation about better than most people. It’s not perfect, but it’s got Ben’s expertise and all the stuff we do around that. It’s pretty good at kind of understanding kind of measurement analysis, comparing what we did versus what we should have done and things like that. So, and then we see people do all sorts of crazy things. I was on on site off on last Friday, a week ago, and the site manager got his phone out and he’s got Claude on his phone, and he said what’s happened on this site since I last visited and what Ray to talk to him about. I was like, that’s quite interesting. So he’s managing multiple sites, hasn’t been this site for a few days, and then came up with five or six things that we saw. This record says we’ve got an issue over here, we haven’t had a decision with the client on underpinning this wall. And I was like, this is crazy. He’s now fully informed, and then he’s going to talk to the team. He’s not replaced those soft skills, but he’s now seen all that data straight away. He’s got 46 variations of all press, spring, understands the people and equipment, and and the notes he’s drafted. That site manager is now the most effective person in that project because we’ve given him a large language model and good records. Nina Young: 32:37 Yeah. Insights and he’s informed and empowered. I mean, it’s literally. Will Doyle: 32:42 And if if you yeah, if you met this person, he’s in his 50s, he is has grown up on site, he knows how to build anything and everything. He’s the most practical human that I’ve ever met with the biggest hands. Yeah, big hands, but with huge hands, about um thick Scottish accent, and he’s got this tiny little iPhone, and now he’s like knows everything about everything. And some people will say, Oh, it should be folks on the site, but now he’s so organized, he he understands, he has that deep knowledge going back to the graduates of people going to profession where they perhaps struggle. That individual I won’t name him because he’s embarrassed by it, he’s so he’s so knowledgeable. Now we’re just giving him a way of accelerating it because he he he he hated the fact of sitting behind a desk and pressing a few keys or uh so but instead he’ll just talk to his phone. That looks a bit weird and wonderful, but he’s really, really effective, and now he’s commercially managing that project. And these QSs must love him. I wish I had 10 of him on all my jobs, so yeah, that that’s kind of quite exciting, and then that’s just a one practical use case of Gavin. Nina Young: 33:47 That’s a really good example, though, of taking someone, like you say, is who wouldn’t want to be desk space, not really into tech necessarily, and you know, in an age group and uh, you know, especially in survey, we we have got an Asian demographic who’s realised as soon as I I I do find that, and are you finding this? Is that there is a lot of fear, skepticism, etc. But once someone actually realizes what it can do for you or your and your work, it completely like it can literally blow their minds. It’s like and then there’s like a light bulb moment. I don’t know, do you see that much? Will Doyle: 34:25 I think the demographic issue is not as bad as people worry about. So we’ve we’ve done a little bit of research on this because I guess all the time I’m worried about an aging workforce, they’re not growing up with these glue to their hands. Um, but they’re actually really good because they’ve done it the hard way, they’ve done the rip the right thing, they know what to say. Right. And the text is just an enabler. And then another use of AI that we is very popular in probably all our personal lives as well is use it for kind of custom support and chatbots, and I need a question answered. We do a little bit of research on who uses our chatbot. Oh, rate to set up our support tickets now are you are answered by Brian, our AI bot, it’s very good. But the people that use it the most when we compare compare their ages is the people that we traditionally would say are not technophobes, I think someone described them yesterday. So we’re now seeing those people love chatting to AI because it’s not embarrassing. No one’s gonna not judging you, you’re not saying, like, oh, why do you not know how the you this iPhone works? I have no idea. Um and they find that interaction with a chatbot really, really kind of easy and not judgmental. And not judgmental. They do that. And I find that’s quite interesting. And we’ve we and we where we see younger demographic are like, did I rather speak to someone? Which again, I if you’d asked me this 12 months ago, I’d be the way around. So we’re not worried about an aging workforce and AI, because I feel like it’s suiting them. I think it’s really suiting the way they want to work. It’s really good to hear. Yeah, yeah, yeah. So when people worry about that, I’m like, look at look at our other customers, look at what we’re seeing over the last it’s always a small sample size, it’s always nine, ten months, maybe. But I I I I I’m seeing that as a really positive way, a really positive use of kind of live side bunch models. Nina Young: 36:06 What’s the um the chatbot that you mentioned? What’s that? What’s that called again? Will Doyle: 36:10 We call him Brian, who uh, what was that? We’ve we person we personify everything because there’s a little bit of kind of great, yeah. Because we we a little bit of psychology in there, but we use an Austrian tool called Gleep, it’s fantastic. It’s a support. That’s why yeah, yeah, yeah. Gleep G-L-E-A-P. Um very small startup, but very AI led as a business. And we are now changing how we do customer support. So traditionally when we do customer support, we we hired lots of people, people really good on the phone and chatting to people, putting someone at ease, explaining something. The last person we hired in support was uh someone with a computer science degree. And now it’s about how do we change our support to be really data heavy so that someone can ping us a message at four o’clock in the morning and Brian doesn’t go to sleep. Brian gives them the right answer that’s just as good on, if not better, than the human. So we we’re changing how we do crystal sort. We’re still doing all that old in-person stuff, but that’s really good. Yeah. When someone’s got a problem, they’ve got a question, what does this mean? Like, how is this score calculated? Like, they don’t need a really nice, happy person, they won’t just want to get an answer. And that’s been a bit different, isn’t it? And now we’re hiring software people and people that are good with data to to optimize that. And that’s gonna be a huge change for us. And we also need less people. So going back to that kind of scary thing about uh employment crisis that we kind of alluded before, we we don’t our support team will probably do double our current user base because of Brian, because of Brian being so efficient. I mean, she’s she’s great for running a business, being efficient, costs rising everywhere you would think, oh, I’ve got to hire more people, that’s a bit scary. An investment. Well, no, we can provide just as good a service now by just feeding Brian F a bit more data. Nina Young: 37:54 Oh, I like that. Brian, right. I’ve done it. Will Doyle: 37:59 So you can talk to him anytime. He’s very good. He’s also he Yeah, so that that that’s the way things are going. I think people what people are hiring for will change. Nina Young: 38:09 That’s and and and there’s been a lot of talk about this across everything, every media, all professions, all over the world. Is there’s the whole talk about, you know, mass job loss, but then there’s also the other side of I think jobs are gonna change fundamentally, it won’t necessarily replace, it’s gonna change the nature. And we’ll be doing jobs and things in the next, I don’t know, easily the next two to five years that didn’t exist before. And one of those is like literally validating and reviewing the outputs of AI, you know, because we have to get better at that. What are your thoughts on that as to where you see it evolving? Will Doyle: 38:49 I think you’re completely right. I think it’s it’s we’re we’re redesigning the factory to use that kind of metaphor again. And if I compare it to our experience in software development, software code, I’m not I’m not a software engineer, so I won’t be to pretend to be too much of an expert on this, but which is it’s traditionally reviewed by another human. I’ve done a bit of work. Can you peer review it? Can we can you check that it’s right? And that’s a kind of a process. Most PRs now are done by AI. There’s tools out there, they’ve spent that they’ve invested hundreds and hundreds of millions of dollars to do that. And we’re seeing that a lot of that kind of checking and review process has been automated because it was just in a bottleneck, it’s becoming more and more of a bottleneck because people are generating so much more code, and so there’s more and more to review to the point where it’s not easy for a human to review a thousand lines. So we’re we’re changing that. So we’re seeing tools actually do some of the review process so that when it does get to a human, it’s when it really needs a human, not someone going through checking someone hasn’t put an extra comma in or or whatever. We want to make sure the quality and consistency of work is correct. And we we use that in gather it in a different principle. Before a human reviews a record, we want an AI to review it. Because we know we want we know that a purchase order should be eight characters, not six. So we’ve got field, basic field guardation, but we know that when we put that purchase order number for delivery, we want the picture of it as well. So have you done that? And it’s very basic, but when we want to compare people’s text to their compulsory fields within Gather, like, well, yeah, I could do that, but if I’ve got a hundred to review every day, wouldn’t it be better if we got automated feedback for the really obvious ones and give that person instant feedback on site? Oh, sorry, I think you made a bit of a mistake there. That used to be eight characters, not six. Just just check it, Ken. Nina Young: 40:37 Yeah, yeah. Will Doyle: 40:37 And that’s at the point, not oh, I’ve reviewed this three days later. Oh, I’ll put that in the bin now. I don’t have no idea what that should be. Maybe it’s a six, seven, I don’t know. So that’s kind of where we’re seeing that little bit of a change. So yeah. Nina Young: 40:50 Yeah, using AI to be like the reviewer, that’s such a good point because I uh one of the things I’ve I’ve said increasingly is around a lot of surveyors that they rely on like peer reviews of their work, their reports and things, and like put it into build your own, put it into AI as well to check before anything else. So you rewrite it, peer review, and then it’s finally looked at the end. You can literally use that to make your work and literally view what you do so much better. Will Doyle: 41:19 And if you look compare that to the coding industry where I was software engineering, we’ve all got great names, like we use code rabbit, don’t know what to call that. Great, but it it it’s it does a very basic function, but it does it really, really well. That’s a bottleneck, that is an issue. And we look at the RICS standard about owning your data, no, and owning your decisions and not just going, oh, the LLN did that. Reviewing, there’s nobody saying that you can’t get someone to automatically review it. And and and that’s kind of where we’re seeing. So I’m quite excited by that. I think there’s gonna be a little bit of kind of change and a little bit of chaos that goes with any sort of change project. But I think the world of kind of construction and infrastructure and surveying is gonna change for the better because all the stuff we hated is probably gonna be automated, and now it’s gonna be really important to do all the stuff that we should be doing, and we maybe perhaps avoiding it because we didn’t quite like talking to that person because they’re a little bit prickly, but we now need to be a bit nicer to them and and spend our time doing that. So yeah, I think errors where we should change it. Nina Young: 42:14 Errors are gonna drop because it’s throughout the whole trade, isn’t it? It’s literally there’s so many people involved, and it’s like if the if everybody starts to adopt AI and use it, reduce errors, that the mundane stuff reduces, drops, drops, drops. There’s less picking up the phone for something really kind of mundane, someone’s missed something because those things are gonna get caught earlier. Will Doyle: 42:36 Exactly. So that’s kind of where we get to. Nina Young: 42:39 Yeah, is there anything else coming towards the the end? I want to make sure that we’ve kind of covered off some of the things we’re going to talk about. What is anything else you’d like to discuss? Will Doyle: 42:48 I think from for me is that I’m not an expert in AI, but I’ve just got to side to go, let’s just embrace it and try and understand the fundamentals and how we can work with it. And there’s lots of better or more well-informed people, and AI-led organizations are doing great work for years before this kind of become the game mainstream. You see those and it’s like planning tools like M Plan are probably the earliest in this and in certain infrastructure sector. But I think what I’d say to people is what I said to when I met Seeker two days ago was this isn’t scary. You can do this within the guardrails and the constraints of you want to use without kind of losing all your data is what people are worried about. And the the barrier to get going is tiny. It it we’re talking $20 a month for a chat GPT subscription. Will it do everything that I’ve talked about out of the books? No. But will it start to make your life a little bit easier? Yes. So when I see organizations not want to give people co-quilite licenses, I hope it’s not because of data security issues, because we can solve those and the costs, well, for an extra £20, £30 a month, we can probably save yourself 40, 50 hours. So that’s the kind of where we’re getting up to. So I think that’s quite exciting. Nina Young: 43:54 I think, and yeah, and I think that’s a really good point to finish on as well, is that literally just try it, get started with it, give it a go. And it’s also, uh as you well know, they’ve got so sophisticated now. Um, you can just have a conversation. I don’t know about you, but I I talk to mine more than anything else rather than typing, because you when you speak, it’s very different. You just naturally you could just talk at it, even for just 10 minutes, and then it will literally transcribe it. And then it’s much more valuable as an input and just talk to it. You don’t need to be this expert prompt engineer. Remember when it first came out, everyone was at prompt engineering courses. I mean, you get to the stage where we won’t even need that, we won’t need that kind of sophistication because the AI is going to guide us. It’s not that scary, you know. You don’t have to be this expert at it, just give it a go. Will Doyle: 44:40 No, exactly. So that’s what that’d be my message. And the the voice is great and it it suits people, just be able to run and discuss. So, yeah, that’s kind of my my kind of overarching message to Will Doyle: 44:50 say. But thank you for having me on this show. It’s a it’s really good to talk to you and um, yeah, hopefully get a good reception and looks it up. Nina Young: 44:57 It’s been brilliant. It’s been brilliant to have you on here and and and share your your thoughts and your insights on everything across, not just um obviously AI, that’s which is uh a topic I’m very biased about, but US generally and the skills and the challenges that have been faced. And again, that sounds brilliant to me. I’ve heard good things as well. Um, but what I’ll do is on show notes, I’ll include links together, I’ll include links that we’ve discussed today. Um, if anyone wants to get in touch with you as well. But yeah, thank you very much for your time today. Will really appreciate it. Will Doyle: 45:28 Thanks again. Thank you. Nina Young: 45:29 Thank you for listening to this is surveying. If you enjoyed this episode, please subscribe and leave a review. It really helps more people discover the podcast and supports the work we’re doing to raise awareness of the profession. You can also join the Surveying Room, the free and independent community from Surveyed UK, bringing surveyors together, breaking down silos, and of course making surveying visible. Just head over to surveyors UK.com to learn more and join today. All the links discussed in today’s episode are included in the show notes.