Episode 20: AI Agents - Accelerating Projects from Chaos to Completion with Greg Lawton

In this captivating episode of the Powering ProjeX podcast, host DK engages in a deep conversation with Greg Lawton, CEO and Co-founder of Nodes & Links, an AI-powered platform revolutionizing project intelligence. 

Drawing from his eclectic background, from stars to schedules to managing large-scale defence programs, Greg offered profound wisdom on the intersection of AI, human ingenuity, and the intricacies of project delivery. His perspectives not only illuminate the challenges of modern project management but also charts a path forward in an era of rapid technological evolution. 

The conversation also touches on the challenges of project delivery, the role of AI in decision-making, and ROI-driven implementations. Lawton also explains how the Nodes & Links platform automates key project management tasks such as schedule analysis, risk prediction, and carbon tracking, making them user-friendly and effective. Lawton emphasizes the need for alignment on strategic objectives and rapid testing for successful integration of the platform into any project organisation.

Moreover, the discussion covers the broader impacts of AI on labour, productivity, scalability, and the role of economic incentives, predicting a future where AI fundamentally reshapes business models and project delivery methodologies. 

The episode concludes with reflections on learning and iterating through failures as crucial elements for continued growth and innovation. 

We extend our sincere gratitude to Greg for generously sharing his time and expertise. His thoughtful reflections are a testament to his commitment to advancing the field, and they provide invaluable guidance for professionals navigating complex endeavours.

Until next time, keep powering your projects from vision to victory, remembering that AI agents aren't here to replace you, they're here to partner with you and elevate you, turning complexity into clarity and delays into deliverables, to keep you powering towards victories!

Chapters: 

00:00 Introduction to AI in Project Management

01:50 Meet Greg Lawton: From Astrophysics to AI

03:25 Greg's Journey in Project Management

07:11 The Role of AI in Project Management

13:25 The Future of AI

Transcript

Daniela Kellett: 0:02

Like what is it about I AI that will make the biggest difference?

Greg Lawton: 0:09

The biggest thing is it will it will enable us to see the world for how it actually is, but not how people perceive it or think it to be.

Speaker 01: 0:22

Imagine a world where your project delays vanish, risks are foreseen before they strike, and your team focuses on innovation instead of paperwork. And what if AI wasn't just a tool, but an autonomous partner accelerating your projects to success? Does that sound exciting? Or are you still skeptical? Either way, stay tuned, as in this episode of Powering ProjeX, we're diving into the Agentic AI revolution that's set to boost project success rates and save thousands of hours in project delivery. We'll explore emerging trends in AI and Gentic AI, the autonomous systems that plan, execute, and adapt to help humans deliver projects faster, smarter, and more successfully. At Powering ProjeX, we champion the growth of project people, empowering every subscriber to lead with distinction, drive impactful outcomes, and turn vision into victory. Through sharp insights, inspiring stories, and energetic conversations, we will celebrate the exceptional people who transform industries and build a better world. Hello, and thanks for tuning in to Powering ProjeX, the podcast where we unlock the future and empower project and business leaders to turn ambitious visions into victorious realities. I'm your host, DK, and today we're thrilled to be chatting with Greg Lawton, CEO and co-founder of Nodes and Links, the AI-powered platform transforming large-scale infrastructure projects. But before we kick off, if you want to join a community that puts your career first, gain global connections, cutting-edge skills, and hear from voices just like yours who are shaping the project management industry, subscribe to the podcast now. Greg Lawton is the CEO and co-founder of Nodes and Links, a project intelligence and analysis platform that applies machine learning and automation to project data. The software outputs actionable insights to help teams simplify project complexity with intelligent automation. Prior to founding Nodes and Links, Greg was an astrophysicist working for BAE systems, managing large defense programs, and advising the board of directors on international strategy. It was during his time at BAE that he realized there was a need for a tool that could simplify the lives of project control professionals by reducing manual tasks. Greg is focused on building a world-class product and an ambitious team who are committed to driving innovation and pushing the boundaries of what is possible with project management technology. Greg, welcome to the show.

Greg Lawton: 3:23

It's a pleasure to be here. Thank you for having me.

Daniela Kellett: 3:25

Greg, tell us what sparked your journey from the stars to schedules.

Greg Lawton: 3:33

So I've never heard it put that way. I thought that that was quite a clever, clever way of putting it. My dissertation was on correctly classifying uh supermassive black holes that were accrued in matter versus stars. Because how do you know if you're looking at a candle that's really close to you or a fire really far away? Basically that question. But you can't touch the stars, and I wanted to do I wanted to touch whatever I was studying. Um so I went into engineering. So I went into nuclear submarine engineering build and design. Yes. Went through that, realized that actually the whilst designing the things was awesome, the thing that caught my eye as the real engineering challenge was the orchestration of how you actually execute the delivery of that engineering scope, which is project management. So I then went into project management, and uh I was actually covering all parts, I was covering contracting, risk, uh, schedules, changes, resourcing, supply chain, everything. But my mind is obviously one of a mathematician and a physicist, so I was looking at it from a very particular direction. Accelerate a number of years, and uh, you know, a lot of theories about advanced algorithms and structured data, of which the schedule is one of the most structured data sets we have. There you get to schedule.

Daniela Kellett: 5:09

Wow, that is an amazing journey. Amazing. And what was it like? So you you're actually working for defense on these projects. I mean, these are multi-year projects, these are not quick to deliver. How did you find it?

Greg Lawton: 5:24

No. Well, there's different it's it they're not one thing. There's lots of different cycles. So you have cycles that are incredibly quick, like days, you have cycles that are months, like project reporting and these kinds of things. You have uh you have long lead procurement cycles, which actually are quite quick, and then you have to wait a long time. There's a lot of these, and then there's a lot of things about purpose, you know, direction. I actually hosted the Australian delegation to the United Kingdom when they were considering their next of tap class submarine, and the policy of the United Kingdom at that point in time was we're we're happy to talk and share a lot of learnings, but you know, we're not in the business of selling these abroad. We you and then I remember the Australians ended up buying the French diesel class submarines, and then they scrapped that when they realized that diesel has a fundamental flaw in that you have to pop above here the water and stick black smoke in the air. It's like, oh, okay. It was a very interesting time.

Daniela Kellett: 6:30

Have you ever read Turn the Ship Around by uh David Marquette? Oh he uh he was um the commander on the Santa Fe, which was uh one particular submarine um that they use in the US um military, and he talks about how he basically changes the culture of the men on board the ship from being um a very fixed mindset, very um, you know, that they felt that they weren't progressing, they were quite demoralized after many years of failure. Um and he manages to turn the culture around and it's an amazing management book. It is definitely worth reading. So, the next question. Uh recently it dawned on me that project delivery success rates have really not improved significantly in many decades. Um, and you have a great reference to the Roman aqueducts on your website, which really made me laugh. You know, and but sadly, research does suggest that we humans inherently don't have really the mental bandwidth to deal with the level of complexity that we're dealing with in today's world. And therefore, the ability to improve our project success rates may not increase over time as it hasn't in many past decades. So, machine intervention and particularly the use of AI in the areas we are not so strong in may be that perfect antidote if we truly want to improve our outcomes. So, if you look back over your own career, what was the pivotal moment in your life when you thought that there must be a better way of doing this? And what was that tipping point that intrinsically gave you the motivation to try?

Greg Lawton: 8:09

Well, that's a very good question. So there wasn't actually a pivotal moment when I thought there must be a better way of doing this. There were thousands of moments when I thought there must be a better way of doing this that kind of boiled up. But there was a pivotal, pivotal moment where I went from being, you know, in my mind, you know, my job was a project manager to my my job was an artist, like to build, create, and do something. And it was uh it was a moment where I had um I had a choice in my career about going direction A or direction B, like uh get the next step of job, get a promotion, kind of go the normal route, or take a chance and and go and pitch hundreds of venture capitalists on a potential future for the industry and work with some of the most successful people in the world in creating an organization that had an ambition to change the world, and that that's the key thing, it's an ambition, it's not a guarantee, it it's not you know, even the right to win, it's the ambition. And it's I I was given those two options, and at my point in my life then, you know, I wasn't married then, I didn't have children. Um, I just had a you know a small a small mortgage on a on a little house. I was like, what's the what's the most I can lose? Like, even if I went absolutely flat broke, I'm actually fine living on the streets, it doesn't bother me, and I've got enough friends that I wouldn't really have to live on the streets anyway, I could crash on a sofa. So it was that moment where I've just decided, okay, I'm just gonna throw everything into inventing things, just inventing something that actually is worthwhile. And the way that I thought about it in my head was that you know, you invent something small, and if it's good enough, people will then give you resources to try and do something a little bit bigger and a little bit bigger and a little bit bigger. And I've just literally followed that path now for seven years. I've done nothing different, it's just I will try and with the team build bigger and bigger and more ambitious things. And if we're we do it right, then we we earn the right to go to the next level of the game.

Daniela Kellett: 10:38

Yeah, this is it. And I think you're absolutely right. I was actually talking to Lia Di Bello this morning. I don't know if you know of her, she's a cognitive scientist over in the US, she lives in San Diego, and she's built a uh learning tool that really plays a trick on the brain. But what it's all about is how do I get you to iterate through failures as quickly as possible so as to be able to rewire your brain to learn and adapt um new skills that you need. And the sooner that you iterate and the more you iterate, the more you fail, the more you learn. Um, and it was been really interesting.

Greg Lawton: 11:15

Yeah, yeah, it's a totally, it's totally a learned skill. Totally learned skill. It's a skill, and part of the skill is a mindset. And the my the mindset is generally this. If I was just saying in bullet bullet point, um number one, the objective of the game is not to win, the objective of the game is to learn. Okay, so if I learn, I've won. Now, like think about that. It's like, okay, if I get fired, I won. I won because I learned something. If I get fired and I don't sit there and go, what happened? Like, give me some feedback, then I've lost. But if I get the feedback, I'm like, cool, I've won. Like who cares? Um, the second the second element is Um your self-image isn't defined by the game you decide to play. And I'm calling it a game for a specific reason, okay? The thing I actually care about is my wife and child. Okay. This thing of going to work is one, something we have to do because we live in capitalist societies, and two, I've chosen a profession that I quite like, and it's to be honest, every day is a hobby to me in this regard. So it it's just who cares? Like, my father did a 40-year career in the banking system, he finished, and one day after the bank, he was like, I I can't remember even what I did. I'm just gonna go and be the world rowing champion now at master's level. And he's like, Okay, so it's that perspective of none of it matters, none of it matters so long as you're holding, you know, your morals and your beliefs constant, and so long as you've got an optimistic and positive persona, and so long as you know you're playing the game of learning and doing better and creating. The world will be as the world will be, and one day I'll be dead and long gone, and no one will remember me, and no one will feel like remember any of the impacts I've had. Cool, so I might as well have as much fun as possible whilst I'm here.

Daniela Kellett: 13:20

Yeah, spot on. Really, really good um perspective. Uh, okay, so let's go back to this. Um, nodes and links co-planner um serves as an AI assistant that guides users through analysis and performs on command, automating workflows on top of existing schedule systems like P6 and Microsoft Project. So, can you explain how that embodies agentic AI in project management today and how it's set to evolve in 2025 and beyond, perhaps shifting from that concept of being a co-pilot to an autonomous autopilot that anticipates our needs and drives even greater improvements in project outcomes like on-time delivery and cost efficiency.

Greg Lawton: 14:07

Oh wow. Do you know what I find quite fun at the moment is the amount of buzzword bingo going around? The the height train, like there's there are definitions as to like what agentic is versus non-agentic and copilot, autopilot, all of this. Okay. Let's cut and say for what it is. This is some marketing wizardry that people are doing here, okay? Like, does does tech do we have agentic AI? Yes, of course we do. There's lots of types of AI we have. We have machine learning, natural language processing, large language models, but there's lots of types. Um I think this like zooming out, like, what is really AI in this context? Um AI is just an automation technology. And I say just an automation technology, it's just an automation technology like the internal combustion engine was just an automation technology for farming. So 200 years ago, 95% of the Western world's population farmed, and 5% were lucky enough to live in big houses and do other things. Today, the other way around, where you have 5% of the population are farmers, and 95% do podcasts and project management and these kinds of things. And so it was an enormous automating technology. If you think about it, and you'll get where I'm going with this in a second, like the internal combustion engine, if you stick it in a tractor, a tractor can plow more dirt than a hundred people with oxen. So one person can do the work of a hundred, which means you can feed a hundred times more people. Now, the reason why AI is just an automation technology, but I'll probably say it's one of it's it's the most important automation technology since computing in the internet, and computing in the internet was the most important since the internal combustion engine, it's because of those ratios. So if you think about it, how many technologies has humanity invented in 5,000 years that automate mind work? Writing, which is remembering stuff. Abacuses. We have invented algebra and things like this, okay? Um we've invented computers, which is automated it's you know, the first computers were called calculators, because the and the first people who sat there doing maths were called calculators, okay? So it's like we've got a few technologies, but not not as many as jet aircraft, you know, cars, motorbikes, cranes, diggers, even like little uh electric like beard trimmers and things like this. Um but now we've created a technology that can self-reason and create to an extent, and it can create audio, written word, visual, and those are basically the three main senses of humanity. So it's like, okay, so we've just invented a technology. Now, if you think about those ratios, it's okay, let's think forward 200 years. Today the equivalent of farmers is general office workers. Okay. Now, you have to split that general office work into many different layers, but how much of the general crap that happens in offices is going to exist in 200 years? So, why have I told you that big long story? Well, if you think about project management, you have things that happen in the real world. Then you have things that happen in the digital slash fake world. Okay. Things that generally go through the digital slash fake world are a lot of documents and emails. Okay. They can be massively automated. And you have orchestration of people. Now, arguably, I could say orchestration of materials and whatever. Materials are orchestrated by people. It's not that you send a signal to a piece of iron and it suddenly floats somewhere. You send a signal to someone to move something. So it's orchestration of people. Now, my big question is will AI eventually start to orchestrate people. So tell people where to be and what job to do and what task card to do. Well, if the communication isn't verbal, and even if it is verbal, if you think it can if it can understand the scope, it can pull out the schedule, it can write the task card, it could understand the resource requirement. You start, you see where the little train of thoughts go in? Okay, so then I'd be like, okay. That's the future, where are we now? We're in the the automation stage. So, for example, the technology we fill automates pretty much every single piece of schedule and risk analysis you'll ever want to do, but keeping humans in a loot. We have machine learning models from learning from all of the project's data. We have uh natural language processing models for starting to investigate carbon analysis. We have a series of large language models and agentic AI that can actually go away and do tasks and work. So, you know, if I wanted to analyze a billion-dollar project and I wanted to know every single change that the general contractor has made in that schedule over the last 12 months, I can not only do the analysis, but I can have a formal claims and change report written by the large language model, you know, 500 pages if I wanted to, and I could have that done within 10 minutes. I'd be like, who else can do that?

Daniela Kellett: 20:02

This is true. And I suppose just moving on from that though, where I was going with the the co-pilot versus the autopilot component is exactly what you're talking about. Because there's a big part of me, like it depending on who I speak to, lots of people have lots of perceptions about what the main problem is with project management. And you know, someone said to me the other day, well, it's it's actually decision making.

Greg Lawton: 20:25

One of the fundamental things that there isn't there isn't a problem with project management. Why do people keep talking about it?

Daniela Kellett: 20:30

Well, I think it's because we have we're hung up on this perception of failure. And the failure is politics.

Greg Lawton: 20:37

That's politics. Like, look if you if you're not looking at it as a social science and you're looking at it like a hard science, and you know, we've got billions and billions of data points in projects all around the world. Projects are pretty damn accurate. It's politics that generally mess it up. And I'll I'll give you an example. So I've got a couple of good friends in Oxa Global projects that publish all of the schedule and cost overum databases. Now they measure those databases from business case approval to project execution. How many projects undergo zero change?

Daniela Kellett: 21:14

None.

Greg Lawton: 21:15

I would say. From the business case concept stage to shovel in the ground. So when all of the general contractors go, this is my plan. We have a grade costs, we've approved contracts. Yeah. From that plan to the very last change that was approved, and then from that very last change to the project execution, okay. The last bit, which is the only stable scope you have, is mirrors makes no difference, 100% accurate. It's like 99-98% accurate. Okay. The change varies between dependent on the project, the geography, and uh the size, and just the state of the customer. So some projects, especially the military drama, have incredibly high because capability changes. Some projects in the um infrastructure world have almost zero changes, like road networks or something, these kinds of things. But it's like, which bit are you actually measuring? Now, what's classed as a failure is generally a politician standing up and saying, We authorized a billion and it costs four. I'd be like, Yeah. You you also were sold a fake dream by people who were paid to get this signed off, and then it went over the fence, and immediately the cost went to three when a proper general contractor looked at this and went, You're not, yeah, this is to do this properly, this is what you're looking at. Oh, and then you added a billion of changes you went through the process. So, actually, if you really analyze it, projects are really accurate and high performing. Politics is not.

Daniela Kellett: 23:04

This is very, very interesting.

Greg Lawton: 23:05

That's a misnomer that people get.

Daniela Kellett: 23:07

I've had exactly that experience that you were just describing as you were saying it. I was like, oh my god, it's all coming back to me. Uh, that's really interesting. So, one of the biggest challenges um is workforce shortage, and you have uh talked about this. So I think you called it the billion person problem in construction. So, how can present agentic AI um technologies like your platform's automated analysis of progress, productivity, and resources address this by empowering human teams, eliminating manual drudgery, or freeing professionals for high-floor? Beautiful question.

Greg Lawton: 23:42

So, actually, you can have a view. I was gonna say predict, but you can't predict. You can have a view on how every single major technology or innovation is going to affect a market purely by looking at the mass and the economics. And there's one thing that you know it's it's nothing is certain in life except death and taxes. I'd add a third to that, which is and economic incentives. People overall, economic incentives predict everything of how economies and people will behave within that economy. Okay, so within that, I said, right, let's think back to the tractor example. Let's just say that it's a 50x automation. Automation. So one, you know, I can automate 50% of the work. Well, a maths equivalent is one divided by that on a 0.5%. I can I can increase the productivity of everyone by 50 times. That's the way. So it's like right. And you see how this is a trickery of maths? It's it's like, oh, I I need a hundred workers, now I need one, or I need 50 workers, now I need one. Well, flip it round. With 50 workers, I can now get enormous, what is it, 2500 units of productivity or whatever it is, from that. So it's like, okay. So the the overarching problem that we have in project management is we do not have enough qualified people. The PMI estimates about 22 million are missing right now. Probably actually a lot higher if you look at how if you look at how much the uh the projects that governments and industries need to do, but they just don't get above the business case mark and these kinds of things. So it's like, okay, let's just take a thought experiment. How would Australia, how would Australia change if we were able to snap our fingers and parachute in two billion very qualified project managers and construction workers tomorrow and let's just if this is just a silly thought experiment. So let's say they don't need housing, they don't need food, and they don't need paying. Well, this I mean, this takes me. What would you do?

Daniela Kellett: 26:08

Well, I'd start with the stuff.

Greg Lawton: 26:09

You'd build everything immediately. Immediately, you'd be like, you know, the planning I'm pretty sure the private would be like, delete the planning department, right? We're not waiting for approvals, just build like and it's like where? The middle, the middle, it's empty. Go there and build stuff, okay? That's what you do. So it's like, okay. And by the way, the GDP of the country would like do this, it would go absolutely vertical, and the bond market in Australia would go, it'd become the richest country in the world in like a year. It'd be that quick. So it's like, okay, that's a really positive thing. We can't do that. But the thought experiment, oh, okay, there's something there, there's something there about massive production capacity that leads to enormous economic wealth. So it's like, okay, how can we start massively increasing the productive capacity? And it's like, okay, well, from what I've said before, AI is one of the technologies that you could start to do that, so long as it's very, very focused. You can't just go, internal combustion engine, therefore farming. You have to go, we need to design this thing called a tractor around it, and also a combined harvester, which is a different engine than a tractor. Yeah. And also, we have to have milling machines that we we need to do. Do you see how there's lots of you've got this thing that turns something, but then you have to design all of the machinery around it. It's the same thing. You've got this thing called AI, and it's like, okay, well, how do I use this to do schedule delay analysis? How do I use this to do risk forecasting? How do I use this to do initial business case creation? And it's like all of those machines are currently being built. People the one of the biggest challenges people have, uh, and this is humanity, not just people, humanity, is is context switching.

Daniela Kellett: 28:09

Yes.

Greg Lawton: 28:10

Everything I'm saying is probably making sense now, but because I've just glued two context together, but generally people don't glue two context together. That that's a big problem, one of the hardest parts of innovation. But it's like if you just look at all of these AIs as just an engine and people are now gluing machines around them to do office work, it's like, okay, that's that that's starting to get pretty extreme. Now, here's the final thing. If you built a tractor, okay. How much effort, how much more effort is it for me to build a thousand tractors than it is to build one tractor?

Daniela Kellett: 28:52

Well, you'd have to build the factory to build the the tractors, so it's that's a it's actually it's actually quite a bit.

Greg Lawton: 28:58

It's actually quite a bit. I'm saying, yeah. Like I still have to I still have to buy the material, I still have to hammer the iron and the steel, I still have to, you know, fit the doors and these kinds of things. Okay. How much effort is it for me to replicate a computer algorithm a thousand times?

Daniela Kellett: 29:16

Probably not as much effort.

Greg Lawton: 29:19

Zero.

Daniela Kellett: 29:19

Yeah.

Greg Lawton: 29:20

Zero copy paste. So what if the scalability of intelligence technology it's far exceeds it's infinite, it's only constrained by our ability to build data centers.

Daniela Kellett: 29:37

Yeah.

Greg Lawton: 29:38

Which is that that explains that by the way. So it's like, okay, coming back this a workforce shortage. If we had an enormous workforce in Australia, GDP would skyrocket and the country would be infinitely more wealthy. Cool. We can build machines to help increase productivity. And we've done that for 200 years. We've actually done it for 5,000 years because you had, you know, the trebuchets and you had the I remember building this water thing from map mirroring what happened in Asians, Egypt in school, you had a lot of machines. Okay. But now it's like, okay, so if 90% of all the work in the Australian economy is office-based work or service-based work in some way, and a large amount of that you can massively increase the productivity with intelligence machines, where are we treading down the path towards?

Daniela Kellett: 30:31

It's a very interesting. Can I just ask you the bit about context switching, which is always really interesting to me because we've all been in those situations where you've got a big project team and you've got sometimes you've got people doing multiple things and they do find it really hard to go from one context to another. Um, and we forget that it's not an easy transition because we are not machines. But what I find interesting about that is project management is a bit like a cube. And I don't know if anyone understands me when I say this, but one thing impacts everything else. If I change my scope, multiple other disciplines are impacted by that one decision. And it is not as easy as just saying, well, I'm only changing scope. I'm actually changing a multitude of other things as a result of that. Do you see that as being a problem with project management? Do you see that happen and in the world?

Greg Lawton: 31:25

It's the complexity of project management. So project management is actually it sits within the the social sciences domain. So if you go to a university, if you do a degree in project management or something, it it's part of the social sciences, which are the soft science component of the science ecosystem. Now, there's hard sciences and the soft sciences. The hard and soft are I actually don't mean better or worse. Hard means that we are assuming the principle of naturalism in the environment. So we're principle of naturalism is there are laws to the universe, and those laws cannot be broken, and those laws are the same in every part of the universe. So maths, physics, chemistry, those are all hard sciences. Soft sciences is that we're looking for correlation on how things behave because we're really analyzing emergent behavior of complex systems. It's really what it's about. So psychology, sociology, business, within business, project management. One of the difficulties of project management is that exactly that it's an ecosystem. And by ecosystem, what I mean is there isn't a single KPI that is this is success, okay? As opposed to, for example, in physics, you're like, what is the mass of something? And you you'd be like, oh, well, maybe it's blue. Yeah, but I don't. That's not the question. The question is what's the mass of this thing? And I've got a whole set of equations that tell me exactly what the mass is at different points of speed and space and whatever it is. In in in projects, you don't. So whilst you have, and that this is why I say it's an ecosystem, because and then within that ecosystem, you have incentive-driven structures because projects are generally bigger than a single person. So it's like, okay, you've got a cost manager that's reporting on cost, and they care about cost and cough that and then you've got a schedule manager who's reporting on schedule. It's like, okay, then you've got okay, do I accelerate the schedule but at higher cost? Do I decelerate the schedule but at higher cost because of liquidated damages? Do I like and this is a complex ecosystem that there are workflows and relationship structures in to navigate, which is like why I say like this whole thing about the biggest misunderstanding that I think the world has about project management is they see it for something that it is not, it is a social science, and what you're doing is signing contracts which are fundamentally promises between humans about a future that they're going to try to create. There's nothing in the laws of physics that that cost is the cost. This is a social science which breeds uncertainty. That is what it is. Like you cannot say, oh, on, you know, this is not minority report. You can't say on Tuesday at exactly 5.01, this exact person is gonna murder this exact doesn't work like that, but you can see you can see harmful trends in the development of someone's you know, extremism that makes the police worried about someone doing something, and then you can see actions they take that make it disproportionately likely that's so it's a probabilistic now projects are the same thing.

Daniela Kellett: 34:56

Absolutely. That's so interesting as you explain it. Okay, drawing on your experience from your upcoming book, The Physics of Project Acceleration, how exactly do tools like yours significantly improve project success rates? For example, is it through faster processing, seeing things that we don't see, return on investment or automated delays and trending analysis? Like, what is it about AI that will make the biggest difference?

Greg Lawton: 35:25

So all of the above, but all of the above are like things in the detail. The biggest thing is it will it will enable us to see the world for how it actually is, not how people perceive it or think it to be. And what I mean by that, I'll tell you two things really about that book. And that book I basically don't mention AI in the entire thing. It's not about AI, it's a it's a it's a cookbook of how you accelerate projects. I think it has all of the mathematical techniques that I have ever come across and could think of inventing to do. And there's two things about it. Number one is I've cheated. It's a it's a it's a book of is a logic substitution. Technically, to accelerate something, you need what's known as um a relative frame, um, which is where Einstein's you know sentence saying everything's relative comes from. Basically, what it is is if I dropped a basketball, how do you know it's falling?

Daniela Kellett: 36:31

Well, it hits the ground.

Greg Lawton: 36:32

You need something behind it where you see that it's moving. Yeah. Otherwise, it's just floating in a space. So you have to have a comparison. Um you cannot compare two projects perfectly. This is true. That you can compare two atoms perfectly, but you same perfect, but you can't compare there's not it's not at the exact same time with the same people, with the same emotional state, with the same intelligence, with the same knowledge, with the same breakfast, same certification, same supply.

Daniela Kellett: 37:05

Same buses, yeah. Yeah, it's not the same.

Greg Lawton: 37:08

And because it's a social science and everything's uncertain, nobody can predict with any degree of s real certainty how a project will behave. Okay. So, if you can't compare it against something and you can't predict it, you can't accelerate it. So, the way that humanity has got round this little conundrum is by creating doing exactly what scientists do, which is to create simplified models of how they think the world works, and then to play with models. So, a schedule is a simplified model based on an assumption of the critical path method of how we think the world will play out. Okay, that's all it's same with the cost, same with a cost uh management structure, same with the risk, same with a risk management plan. Yeah, the they're all they're all just models of how we think the world. By the way, same with general relativity. Do we do we exactly know how gravity works? Oh, of course we don't. But we have a model that seems to be weirdly accurate in predicting what will happen. So the chances are the universe follows rules probably similar to something like that. Yeah, that that's the whole basis of science, is we create models that are summaries of how we think the universe's laws work and then we test them.

Daniela Kellett: 38:41

Yep.

Greg Lawton: 38:42

Okay. So if we if we go back to the book and projects, what the book is, is all of the maths techniques or the techniques that you can do to the model to make it go faster.

Daniela Kellett: 38:57

Wow. Okay.

Greg Lawton: 38:58

So that's number one, it's it's a substitution. It's called a substitution fallacy, but the whole world assumes that that's how the way the world works. So um we'll go with it.

Speaker 01: 39:09

Is that another fallacy?

Greg Lawton: 39:12

Yeah. Um, so when I come back, when you the question was, how can AI make the biggest difference? It's things like, okay, that it seeing the world how it is, as in, it's not a real world, it's the state world that we've invented called a model. Um the second is because of that, it's a flip. The question is not what can I practically do on a project? Let's calculate if the model says it goes faster, it's the other way around. It's let's calculate all the stuff that theoretically makes the model go faster, and then go and ask people on the ground if it's possible. So AI is a massive automation technology. How can AI make a biggest difference here? AI can run every theoretical acceleration methodology that is that is physically possible within the rules of the model that we've developed, it can cherry-pick a few, and then it can take it to the project team and go, is this doable? And like it doesn't know, it doesn't know if it's doable, but it can ask them, and if they go, Yeah, yeah, that's probably doable, sweet. The model says we go faster, and by by the rules that all of humanity has currently agreed on, that makes a project faster.

Daniela Kellett: 40:32

Wow.

Greg Lawton: 40:33

So, like there's some really, really deep thinking you have to do in to really understand this area. But once you get there, it kind of all makes sense.

Daniela Kellett: 40:46

So this is interesting because I'm gonna ask you a question a little bit later on, which is about processes. And if a machine is sort of saying, well, this is the maths that we can use to try and work out what the best way is to deliver this project to move as quickly as possible. I assume that that's what you mean by this. No what that could also potentially do is say, uh, all these processes you've been using in the past, you didn't really need any of that. Uh how about we do it this way and tell me, is that possible?

Greg Lawton: 41:19

If if you if you have models big enough, yes, of course it's possible. Because if you think about what a what is a process, okay? A process is um a codified, repeatable way of doing the same thing, either making something or choosing something in some regard. And a process sits on a scale from I'm just making it up as I go along, which is known as creation, all the way to systematization, which is essentially codifying processes in technology. So computer says no. It's like you have to click. So for example, Amazon. Does Amazon have a payment process or does it have a payment um systematization? You can't no other way. Because you you there's no other way, you can't order a product without paying. So technically the whole thing is a process, but you have creation, which is one offs, then you have manual repetition, which is far easier to spin up, but has errors within it, and then you have systematization, which is much harder to spin up, but theoretically has a lot less errors in, and you get colonies of scale. Now, if you think about a project as unique activities, and then a series of repetitive processes that are either creation, correct, manual, or systematized, then that you can start to use these intelligence machines. You would need an intelligence machine wrapped in a shell of it's a process chooser. They can start to look at data and start to optimize processes in uncertain environments, just like we've optimized processes in more certain environments, which is production-wise.

Daniela Kellett: 43:31

Absolutely.

Greg Lawton: 43:32

So, absolutely, it wouldn't be a master, you'd have an AI machine specifically for process choice. Wow.

Daniela Kellett: 43:42

I love this. In high-stakes environments, interdependencies often lead to hidden delays. Um, how does nodes and links platform currently use advanced AI in risk prediction, acceleration, and carbon analysis to forecast and mitigate those risks? And I'm really interested in this because I am a big believer that we struggle to make complex decisions. And I really want to understand how do these tools help us improve our decision-making capability. You know, you and I have both been on projects where decisions are sometimes delayed, which cause all sorts of issues, or we just don't know the best, we just don't know the best decision to make, whether it be because the person making it doesn't have enough knowledge, doesn't trust the information that's coming towards them. So many different reasons, but I want to understand how the platforms and the AIs will help us with that.

Greg Lawton: 44:36

Got it. So I'm gonna say something to start with that will be incredibly surprising given that I'm the CEO of a company that builds fundamental AI for project management. Um the system is exactly the same as people have been using for 20 years. And I say that in the user interface and the choices that people make. And the reason for that is the goal here is not stand on ceremony and say, we have some pretty ideas, these are better than everyone. Shut up. Okay. The goal here is to genuinely help improve the productivity of knowledgeable people, and you massively do that by making it unbelievably easy to adopt and start using. If you need, if you need, you know, to the extreme level, if you need more than a one-hour training course, you've already lost 95% of the people who will use your system.

Daniela Kellett: 45:37

My judgment of tools will have to be able to use it within five minutes. If I can't work out how to use it straight out of the box, uh, that's it. I don't want to read anything.

Greg Lawton: 45:45

Exactly. That's overall. Exactly. And and here's the here's the thing: that's how it should be. That's how it should be. Like the there's a lot of there's a lot of people I've spoken to which are more academically minded, which are like, oh, we can build better models for this and that and that. And I'm like, irrelevant. Irrelevant, it doesn't matter. Why? Because nobody's gonna use it, so you're gonna have zero impact on the world. You'll sit there in a castle with a brain fart that might be technically true, but implementably a piece of shit.

Daniela Kellett: 46:14

Yeah, yeah.

Greg Lawton: 46:15

Okay, so let's separate those three. So uh so it's exactly the same, and you can get up and running and using it within 30 seconds. So when you log in, it goes upload a schedule. Boom, I've done everything. An AI agent, what do you want to see? I want to see this. There it is. Do you like it? Yeah, I do. Can I see this too? Boom, there it like just instantly, and a lot of the stuff is automating things that are not AI native, NS code. Like, there's no AI in that, it's just a normal algorithm. You so but the AI knows to go and use that algorithm to do that thing to get this thing and then to present it on this graph in this way, and then it knows how to describe it with this text. So it's like that kind of thing. So, number one, it's not any different at the org. That then comes on to a very interesting topic, which is the more advanced stuff like acceleration, risk management, and these kind of things, okay? Again, I'll come to my point. Theoretically, you might be correct, but if no one knows how to use it, no one trusts it, and people speaking frankly can't be asked, it's worthless. It's worth less than zero because you've spent money making it. Okay? It's it's worth negative to humanity. You you you would have done better for humanity not doing it. So if you take that view, like the risk, and by the way, we've gone beyond and made those mistakes. Like, we've invented, we've invented like very advanced automated machine learning models that can learn from loads of stuff and predict, but unless they're to an extent explainable, no occurs. So, like what we've got is the normal QSRA, Monte Carlo simulation processes, but you have choices to use AI. So you can choose to use machine learning, but you can also choose to tell it to shut up. You can also use large language models to choose to write the report. You can also choose to write it yourself. It is it is the user's choice, as as you know, people would say, it is the dealer's choice in the casino. Choose what you want to do. And you know, there's a scale. I could go down a route of some very advanced theories that we have, but it's irrelevant at this point. Projects need to evolve at the pace that they will naturally evolve at. And one time in the future, those theories may it may be time for them to come to light. But right now, they are worth less to humanity.

Daniela Kellett: 49:01

Yeah. So are you saying that people don't use the platform for um tell me what decision you would make, what you think would be the right way to progress? Do they use it in that way?

Greg Lawton: 49:14

Progressive what? Not with a decision. Like you showed me an S-curve, I can see that there's a problem. Um, what do you recommend I do? Well, it can make recommendations, but it can't enforce those recommendations. There's no AI robot with a gun to your head. So, like, you can go, that's nice. I'm going for a coffee and I'm gonna think about the songs I listen to other. Like, it's you can just totally ignore what it's saying. Yeah, and it's got no emotion, it's not getting frustrated. It's like just like so like the the platform specifically, you know, it automates all of the data integrity analysis and it can identify how to fix certain schedules and these kinds of things. Automation. Um, it does all of progress tracking, resource tracking, advalue management, all of that kind of stuff. Really big things that people use it for is change tracking and change management. So, you know, if a if a general contractor or a client or whatever changes something, rather than taking two weeks to analyze and understand, 30 seconds done. Flat. I know everything that's happened. Uh, you've got delay and forensic analysis, you've got forecasting, you've got Monte Carlo risk simulations, you've got carbon analysis, you've got portfolio analysis. Is the whole to my analysis before it's a lot of machines, a lot of machines just packaged into the same place. And the platform is the equivalent of you opening an office door and talking to people and there's loads of people sitting at desks, and you're like, delay person, yeah, do me that thing, and they're like, Yeah, sure. All right.

Daniela Kellett: 50:39

How are you going with all the things? Yeah, change person, do me that thing.

Greg Lawton: 50:42

Yeah, yeah, that's basically what it is. And yeah, of course, it's like, I think you should do this, I think you should do that, but you can't force anything to happen.

Daniela Kellett: 50:49

And I found that really interesting. The more research I do into sort of, you know, I'm learning, I'm not like you, but um, how AIs are structured and multi-agent AIs and multitasks and so forth. It actually surprised me at of how much it looks like a just a project team or a team in any business. I've got a manager, I've got all these workers, they all have to hand off to each other, these guys work together or a brain correct, correct.

Greg Lawton: 51:17

Your brain is at one thing, like your personality isn't one thing, your personality is many things, as my wife will attest to me. If I'm hungry, my personality is very different than if I'm like had a good night's sleep and a great breakfast. It's like there's different parts of your brain, and your brain hands off signals to other parts of your brain to do things, it's the same way. It's a complex system, it's a social science. Yeah, and I loved the way that they were describing it that you know, you are better off getting AIs to do very small tasks and then getting multiples to work together. It's like, my god, this is just like humanity, it's like we are recreating ourselves.

Daniela Kellett: 51:54

That's actually that's actually in my mind how the general intelligence that everyone's kind of aiming towards will actually happen. Because if you think about it, so I'll give you an example from our platform, okay? I can have I've got an AI that could understand human language and translate it into machine code. Okay, so you say I want to see an S curve, and it's like I've then got an AI that can go and pick the right algorithm from a bunch of algorithms of how to produce it. And that is the S-curve algorithm and the presenter algorithm, the graph. Okay. Then I've got a piece of AI that can take that data, understand the concept, and write a description. Okay, so I've got my description. Now you could go, okay, this is great. I want a progress report. I've got the same AI at the start, understands that, and sends it to an AI that's for reporting. Now the report AI gets that and goes right. It needs to be on A4 paper, it needs to do this. Like our system's already got hundreds of little AI machines to give the to give the perception of grander intelligence. Key thing about an AI is you can you don't have to program in every response it could possibly give when you create it. Okay? That that's the key difference between, for example, just the deterministic code. It it's a whole set of these little machines that then result in the perception of greater intelligence.

Greg Lawton: 53:22

It's just like building a building, isn't it? It's all these little components, and then we stand back and there's the perception of a building. Same thing. Um, all right. So looking ahead, in which areas such as risk prediction or carbon tracking do you see or foresee the most significant gains and how might technology improve over time to compound savings beyond your current 10-hour um time saving per person benchmark?

Daniela Kellett: 53:50

Yeah, that's a really easy thing to that's a really easy thing to achieve. Like that we we we picked 10 hours because 100% of the time we can think about chat GPT, it's like write a report, but ours is very geared to what you're doing. It's like, do you do this process? Yes. How long does it take you? 20 hours. Cool. I just did it in five minutes.

Greg Lawton: 54:08

It's like that was an easy win. That was just a very clever pick by our marketing department.

Daniela Kellett: 54:14

Well, I thought it was so clever. Um, all right, but going back to the question, what were we saying?

Greg Lawton: 54:20

Um what's the most significant gains? Okay, I'm I'm I'm gonna go on layers here because gains to poo is the question. And what is a gain? So look, again, it's an automation technology. That's all it is, okay. The the only real cost that exists in projects is people.

Daniela Kellett: 54:50

Yeah, and materials, labor.

Greg Lawton: 54:53

Technically materials, but materials are people at some point in the supply chain. Like that's and financing. Oh yeah. Where's the money coming from? Those are those are actually the only cost and you might go, well, what about risk? Well, what do you spend it on? People, materials, or more financing. Well, what about loss revenue? Okay. Well, loss revenue is just the inverse of either people, materials, or some kind of financing. It's a financing problem, loss revenue normally, because it changes the the net present value calculation, which changes the loan rates, blah, blah, blah. Like, that's why it matters, okay? So it's finance. It everything is just one of those, one of those things. So it's like, okay, AI is a massive automation technology. Your core costs are people, materials, financing. AI increases the productivity of people by a hundred times. Let's just say that. Your per unit labour cost crashes. Because of that, your your ROI calculations on your projects explode, which makes financing easier.

Daniela Kellett: 56:12

Yeah.

Greg Lawton: 56:12

And materials is people along the line someway, but it's also a coordination thing. Now I could talk about AI doing certain tasks better than people. Fine, whatever. It's it's just a di it's this it's the other side of the same cloth. Being more productive can be measured as just not getting the answer wrong, as well as just doing more words per hour. Like it's the same thing, basically. Okay. Now, that's one. But then it's like, okay, I'm I'm changing the fundamental economics of a market. Business models sit on top of markets. How do business models change? Well, most of the consultants in Australia are per hour or per day.

Daniela Kellett: 56:56

Correct.

Greg Lawton: 56:57

What happens when people can produce 10 or 100 times more in a day?

Daniela Kellett: 57:03

Yeah.

Greg Lawton: 57:04

Open question. General contractors. What happens when you can have AI, for example, do all of the legal analysis work and submit all of the claims that you're entitled to every single time. Like, it becomes a very interesting place. So at the base level, these things happen, but the biggest change that I see happen happening is power laws in business models. And I'll give you one final number here to like make you really think. I don't know what the eventful answer is, but the average sh the average revenue multiple for a pure software company on the US stock exchanges right now is about high growth, 10 times revenue. So if you have a hundred million of revenue, you're valued a billion dollars. The average multiple on a mainly labour-based business, like a consultancy, is 0.8 times revenue. Not all dollars are made the same, not all money is the same colour. So, picture this what if a tech consultancy came in and just charged 10% of what a consultancy charges for exactly the same service.

Daniela Kellett: 58:34

Correct.

Greg Lawton: 58:35

And by the way, they still have some people there.

Daniela Kellett: 58:37

Yeah, just checking, validating.

Greg Lawton: 58:39

That's the genuine thing. And the market will still value them, ex- They'll value them more highly, is what they'll do. So what we're going to see in the next 10 to 20 years, it's not gonna be media, it's gonna be 10 to 20 because we're in a complex environment, is the absolute death of so many companies that just aren't its natural selection, just aren't evolving to understand how economics of this market are changing. And you're going to see correlatory the growth because if the overall market is growing because of the pro the increased production, people are dying, somebody's getting a lot bigger. So you're going to see absolute juggernauts emerge that cannot be sought. And I'll give you the exact example in defense Palantir. Trillion dollar company. It made something like 200 million of revenue of profit last year. Trillion dollar company. And I'd be like, okay, what multiples that?

Daniela Kellett: 59:54

I don't know. Do you know? Well, it's I think it's about I can't remember, it's like 500 or Something like that, but it it they're getting a bit crazy, but I'm like, no construction company is valued that way.

Greg Lawton: 1:00:06

Yeah, that's crazy. And what's doing what's driving them? Palantir scalability. Scalability. Wow. Scalability. It's the fact that the fact that they've invested billions of dollars over decades in building out all of these AI machines specifically for the use cases that they do. And right now, if they go head to head with a normal company, you know, the normal companies bring in a stick and they're bringing an entire battle chip group. And it's like, all right, they'd be like, I'll tell you what, we're expensive. But we'll just undercut you by 3,000%. How about that? And and by the way, they're still making profit. And the other company's like, uh, what?

Daniela Kellett: 1:00:49

Correct, correct. Did you see all of the graphs that were on LinkedIn? I'm sure you don't have the time for that kind of stuff. But there were graphs about Gartner and their share price dropping off. And I didn't know if they were real or not, but it was like this is already having an effect.

Greg Lawton: 1:01:06

Yeah, well, if you think about search, you've got you've you've got like you've got the infrastructure, the infrastructure, artificial intelligence, physics, subplexity, open AI, those kinds of ones. And by infrastructure, what I mean is they're making the core generalist models that cut across every single industry. So you can apply their models to lawyers as much as you can apply it to newsreaders, like uh generalist models. A company like Nodes and Links is a vertically specific AI company because we have to make things for a very specific context with a very high bar of errors. So, like a general model can make errors and it doesn't really matter too much, but you can't make errors on things like project acceleration and these kinds of things. But winding uh winding it back, people like Gartner, their business model was fundamentally predicated on the solving the problem of information asymmetry. The fact that they knew more than you about a about a marketplace or a specific place. Okay, well, now we've built intelligence machines that can multiply the productivity of one person in that regard by a million. So a person with very low skill can investigate a market as thoroughly, or maybe a little less thoroughly, than an entire team of highly paid professionals. And if I wanted a membership to garner, because they pitched me quite a few times, it's 40 grand a year.

Daniela Kellett: 1:02:34

It's very expensive.

Greg Lawton: 1:02:35

40 grand. Or I can just go to Chat GPT and I can ask. So it's like I don't care anymore.

Daniela Kellett: 1:02:41

Correct. Correct. And even if it's just that little bit, you know, error prone or whatever, you get the general gist enough information to be able to make the kind of decisions that you need to make. So I agree with you wholeheartedly. Uh, another one emerging trends show agentic AI enabling predictive analytics, real-time tracking, and multi-step planning, making AI more of a partner and an advisor in project delivery. So you also made a comment in one of your videos on AI stating that AI will never take over project delivery as it can't perform the engagement workshops and responsibility allocation like humans do. So, do you ever see AI becoming a project management partner at least?

Greg Lawton: 1:03:30

Uh yes, totally. Well, a question here is the computer that you're using to talk to me your partner?

Daniela Kellett: 1:03:39

Yeah, in a way. I couldn't function without it.

Greg Lawton: 1:03:42

Context. Context, okay. People would be like, okay, it's about the frame in which you see the world. If you go back 50 years, people will be having exactly this conversation about computers. They'd be like, Do you see a computer being used in all aspects of the work in design and every office? And be like, yes, I do see it here. It's the same, it's a generalist technology that like offers advantages on certain things, and it's like, okay, which is computing, visualization, and communication. So it's like the same thing. It's like, okay, you've got something that can talk, draw, yeah, make videos, and reason. A question, and it's I I do this in some of the presentations I do. I go, all right, imagine you're in 1950 before the invention of the computer, hands up. How many of you, as project managers, because project managers existed back then? How many of you would have used a computer? Nothing. And I was like, how many of you today use a computer in your job? 100%. And I was like, okay, so at some point it went from zero to one percent. And at some point it went from 99 to 100. And I'm like, okay, hands up who uses AI now? And a couple of people put their hands up and I'm like, so we've gone past the zero to one percent stage. And I'm like, okay, the only question now is when is the 99 to 100?

Daniela Kellett: 1:05:10

And what did I I saw a thing the other day, it took 13 years to get from horse and cart to pretty much full vehicles down, you know, Washington Boulevard or whatever it was, and there was a it'll be about the same, it'll be about the same for project management.

Greg Lawton: 1:05:25

It'll be it'll be less if you're just talking about do you use any kind of AI? It'll be like we're talking single-digit years. But if you're talking about integrated AI into every single process within project management, you're talking 10-15.

Daniela Kellett: 1:05:39

Another question for you, which is a little bit off script, but when we talk about that whole concept of AI, right, it's it is a little bit um there's a bit of disillusionment here because AI can be embedded in things that we I don't even know that it's embedded and that I'm using it, right? So the reality is we could be well past 1%, we could be at 45% by now. And I really wouldn't know because it is embedded in the tools that I'm using every day. But I want to understand, do you think that AI will evolve more at an individual level or at a commercial level? And the reason I say this is because there is so much scope now with the tools that we have with the no-code, low code for PMs to build their own AIs. And I sort of, instead of rocking up with my briefcase, I rock up with my computer and I've got all of my AIs that I use for project delivery, and I've built them myself, and they are mine, and I take them with me from job to job, from gig to gig. Um, and then I also use platforms like nodes and links, and I can run, you know, bigger schedule analysis and all of that kind of stuff. Do you see that happening? Do you see that being our life?

Greg Lawton: 1:06:47

Um Right. So one principle in innovation is that nothing is ever new. It's just a different concept, okay? Or applied in a different place. So I use the analogy of the combustion engine to kind of put a basis of logic of to what AI would do. Okay. It's just automate, nothing's ever new. And in in this question, I'd say the same. Nothing is ever new. Okay. People have had Excel for 30 odd years. Do you bring Excel models with you when you go to a job? Or people have had Microsoft Word. Do you bring a reporting template with you to a job? There you go. So to some extent, to some extent, yes, people will take their own AI. But to another extent, there'll be corporate policies, and especially around information security, that will turn around and go, no, and there'll also be corporate incentives where a company is more concerned with maximizing the profit of the company than allowing each and every person to be their own little artist and creative endeavor. It's like there'll be balances to these things. So they might like they might let you bring in a couple of reports and they might let you use a general AI for stuff, you know, research. But when it comes to like, and I know company financials, are they gonna let you put their their internal profit measures through some they'll be like, no, no, no, you're fired if you do that.

Daniela Kellett: 1:08:16

But I think it will be like So it'll be a balance. It'll be like the template model, right? It'll be like if you want to do financial reporting for the company, we hold all of that gold-plated data. We will give that to you, tell us what you want in your report, and we'll be fine. Um, and then there'll be other things. Well, this is my my tool bag of uh templates that I take with me and I use.

Greg Lawton: 1:08:36

A perfect example of this is documentation of like a project management plan, a PMP. Okay. Most PMPs, you can just delete the name and insert a new name, and it's like this is how we're gonna deliver the project, these are the processes, blah, and you might tweak a couple. Cool. Well, you might have a PMP generating AI that asks you some specifics and generates you a document. That might be yours, and then as soon as you bring it into the company, under your contract, all created works that you create becomes the company's IP and it becomes a document. And it's like, okay, like that.

Daniela Kellett: 1:09:10

That would work. All right. Um, for the future, where do you see the biggest advancements in outcomes? Will it be enhanced predictability or sustainability or something else as AI evolves?

Greg Lawton: 1:09:25

It'll be whatever humanity cares about, to be honest. Like, could we have a massive improvement in sustainability? Yeah, if people cared. Um, and by cared, what I mean is you you actually put a dollar value to it. Like, again, incentives rule all that you are everything's certain in well, the only three things certain in life are death, taxes, and economic incentives. So it's like if you said for each ton of carbon you reduce, I'll pay you a billion dollars, there'd be no carbon tomorrow.

Daniela Kellett: 1:09:55

Yeah, correct.

Greg Lawton: 1:09:55

Like it'd be done. So all I'd say is it just depends what people care about. And this is the thing, which is we all exist in a in the Western world, in a free market economy. An economy starts with a person and ends with a person. You don't start with a machine or end with it. There's no such thing as a machine market participant. Now, can you have people and machines in this complicated mess in the middle? Of course you can, but it starts and ends, demand and supply. If people demand it, that's generally what will happen, so long as they demand it through incentives, i.e., I will buy it or I will not spend money with you if you do that.

Daniela Kellett: 1:10:41

Yeah, correct. Correct. It's very true. Uh, based on your mission to make AI in project controls a reality, and I think it already is a reality, to be honest. What's well past that now? We're past that point now. Um what's your advice to business leaders on leveraging present technologies that demonstrate a clear return on investment from day one? And how do you envisage that the technology progressing over time over the next few years? Is it through better feedback loops, expanded global scalability, or gains in areas like stakeholder engagement and claims success, which you've brought up?

Greg Lawton: 1:11:21

Um, so so I can I can't give a global view, but I can just answer how we do it inside nodes and links. So on the first one, which is um demonstrating clear IOI for day one, we've got we've got a few rules, probably rules of thumb in the company. So number one is if you can't if you can't measure ROI within four minutes, just stop wasting the person's time. Like that fast. Okay. And it's things like okay, open up the system, give me a schedule, find me every single change that's happened, and write me a report. Pum, done, two minutes. Is this your report? Oh my god, that and then number two, which is related to that, which is measure ROI by how open their mouth is. So if they're like right, there you go. That's the ROI. Is it everything else? All the numbers about you know, automation, de-risking, they all just work out. But the you know, the real ROI is an emotional ROI. If people are like, Yeah, like I've spent the last 25 years manually doing that, and you've just you've just done it for me in 30 seconds, it's and generally people are like, I I honestly can't believe like if you asked me 10 years ago how my life would progress, it wouldn't be this. This is not what I would be saying.

Daniela Kellett: 1:12:36

Yeah.

Greg Lawton: 1:12:37

So that's kind of how we how we provision that. Then how do you envisage technology progressing? So uh there's this amazing quote, I forget who it's from, but it's from a very successful entrepreneur, which is like, what's the one piece of advice you give to young entrepreneurs to stop them failing in their business? And the quote is something like Nobody cares what your product vision is, they only care about what they have to do. So, so we in those in links, we actually only build what people ask for. Now they can ask for it directly. So I want this automation, but they can also ask for it indirectly by us being like, where did you spend all of your time last Thursday? I spent it doing this being claims report. It was a nightmare. Really? How many of those do you do? Oh, I do like 30 of them a year. How long do they tell you? Oh, yeah, yeah, this.

Daniela Kellett: 1:13:34

And it's just like I wrote a magazine.

Greg Lawton: 1:13:36

Yeah, and then you're like then you're like, okay, and you're like, Do you know that thing that made you open your gob a couple like a little bit ago? Yeah, yeah. If I could do the same for claims, would you open your gob just as wide?

Daniela Kellett: 1:13:49

Wider, wider.

Greg Lawton: 1:13:50

And you're like, okay, I'll give it a go. Alright, cool. That's so it it is about that constant engage. Not yes, there's some grand vision at the top of it, but the tactical execution and and and by the way, this is the you breed your own words inside the company. The thing that makes nodes and links unique is that we are AI thirst at heart. Now that doesn't mean you use AI for every single job. What it means is we are pushing the absolute boundaries, both with our product but also our internal business operation, into how massively productive we can be. So if our if our um engineering team isn't experimenting with the with the very latest technology to push their production of code fast as possible and their quality checking and everything, what are we doing? If our sales team isn't using AI in some way to help structure all of our reports, to help identify where we could serve customers better, what are they doing? So it's the same thing. And the thing that really makes us unique in that is how much the flywheel has spun over the years. Because you don't you don't just do that and that's it. It's like a sport, you get better and better and better and better and better at it. And now we're at the point where our team can produce more than a team about five times the size of it on the regular.

Daniela Kellett: 1:15:29

And this is interesting. That's where I work in IT departments normally, and one of the things that I see is these processes that have become so enormous and so rigid that we cannot cycle through anything quickly. And there is absolutely no appetite to improve processes at all. You know, there's a cursory discussion about it, but it never actually manifests. And what you see is invariably the size of what we release decreases year on year, and the amount of work that we have to put into just getting through one release every every month is just extraordinary, and how much stuff gets pulled out of the release.

Greg Lawton: 1:16:10

All right, I'll I'll come back to the thing. There's only three things certainly in life death taxes and economic incentives. If I came into your business and went, I will pay everyone a million dollars if you are able to reduce the processes within four hours down to 10% of what they are now, I bet the processes would be down would be cut by 90%.

Daniela Kellett: 1:16:30

That's true.

Greg Lawton: 1:16:31

And yet it It's about want.

Daniela Kellett: 1:16:33

And it's interesting what you say because when you really think about a lot of organizations that I work for, a lot of the people that are hired at that management level and leadership level are all about compliance to process right.

Greg Lawton: 1:16:47

By the way, is why the business which is why the business cycle exists. So why do startups exist? Because arguably you'd be like, well, all of the big companies have more resources, and technically all of the people the reason why startups exist is because businesses go into kind of a high the nation mode and they try to optimize for the cart and horse. And what you have is incredibly ambitious human beings who don't want anything to do with that, who just want to invent and create and go the next level. And what happens is businesses grow and either destroy those historic businesses, blockbuster, those kinds of things, or that big business eventually goes, Here's a lot of money, please give us yourself, and then and then really what they do is just do an internal switcheroo, so they actually are the new company, just badged under the old name, and they just cycle like that. Like the whole people, the competence of people is based on their skill set, their ambition, their discipline, and opportunity. Okay, and there are groups of people who are incredibly different in skill set and ambition. Each you know, each section can have incredible discipline, but if you sprinkle opportunity on these people, you've got something real special. And that is what makes billionaires in Silicon Valley.

Daniela Kellett: 1:18:22

Yeah, absolutely, absolutely. God, I could talk about this all day. Alright, looking forward, how do you see Agentik AI reshaping delivery methodologies and democratizing project delivery in 2025 and beyond?

Greg Lawton: 1:18:38

So I've got maybe a shorter answer for this, not a particular answer, okay? The current education on project delivery methodologies is incredibly poor in my opinion. It won't be a genetic AI that reshapes this, it will be people properly thinking about how delivery should be done. Now, can automation technologies enable people to come up with ideas and to validate those ideas with data far faster? Of course they can. But fundamentally, the the education and the humanities capacity for understanding delivery methodologies is far below what it could be. And I'll give you an exact example of this. In production line operations, okay, if you've got two conveyor belts entering the site and they have different rates of component production, and you want two conveyor belts exiting the other side with equal levels of component production, you need a balancer. You need something that shifts units from the bit with more stuff on it to the side with less stuff, and then it needs to stop switching when they're equal, okay? In projects, project a lot of projects trigger when the design is not complete. Okay. The design team will be told finish, get the design to the required level and do it fast. The way you get design to a required level quickly is through massive iteration. You have to iterate as fast as possible. Okay? So so if I want design to be complete as soon as possible, I need iteration speed to go towards infinity. The execution team. They're generally like managed on predictability. Okay. Predictability and efficiency. Predictability and efficiency are about repetition of the same thing. About going up the learning curve and getting those economies of scale and economies of learning. Okay. You want the iteration in what you're doing to be zero.

Daniela Kellett: 1:21:25

Yeah.

Greg Lawton: 1:21:27

Why the fuck are people confused when projects that have an iterative delivery cycle and a predictability-focused execution cycle don't meet very well and end up overrunning? Because you built the world's most idiotic machine. Like, let's not be surprised at this when this does not conform with a standard line. Okay? Now there's loads of stuff like this. It's like, okay, well, what about off-site manufacturing? What about wanting acceleration, but HR having established massive, massive gated union contracts that allow no overtime?

Daniela Kellett: 1:22:06

Correct.

Greg Lawton: 1:22:07

It's like, okay, so like in this point, I'm like, there's an entire there's an entire science around a grand project strategy that needs to be more refined to enable people to make these decisions. Now, the theories of that can be suggested and rapidly validated by AI using data. That's what's needed in my mind.

Daniela Kellett: 1:22:32

Okay, that is very interesting. Um, I'm gonna go to the last question now. For our listeners that are eager to action and, you know, to have a look at nodes and links and actually use it, but they need to build a business case for how the company can measure the success of implementing a platform like yours. What would be the best way to accelerate or to articulate the potential for acceleration and wins?

Greg Lawton: 1:22:58

Uh so so the business case is generally just a process that's based on general ROI. ROI is super easy. You just feel like, oh, save this time, extrapolate out to everyone. Look, enormous savings, cool. Okay. Get us off the lag. Um the real thing is about the um the big themes, strategic themes that the organization wants to pursue. A lot of them have AI now, and a lot of them have implementing AI. Now there's two forms of that. One is implementing something that wasn't specifically designed for your company and trying to find a use case and fit it in more. It's so much harder than you think. Like, even the chat GPT is like, oh god, like what is an API? How do I do this? Like, how do I reinfense data? Like, it's actually uh very hard. Um, and then the other is just very vertically specific technologies like like ours, where it's like, I'm going to do this thing, so I'm not going to do HR. I'm not gonna do it, okay? But if you want a delay report, I'm your guy. So it's like very specific. So that's number one. Number two is and separate the business case into two stages, and it comes back to the point I made right at the start. Stage one is learn, stage two is scale, okay? And this is actually where I see the race happening in our sector right now, is okay, let's imagine that AI is the thing, like that all consultancies or all GCs will be using AI. All right. But some point you have to learn because you can't just go, I want to implement this everywhere. You're like, no, I need to really understand how people react, what the change process needs to be, blah, blah, blah. Okay? And then you need to scale. Is it better to have time to learn and to not be really pressed? Or is it better to try and scale at the exact same time as learning because all your competitors have just started to do it?

Daniela Kellett: 1:25:08

Well, you want to iterate learning as quickly as you can, but at the same time, you want to do that, yeah? To make foolish mistakes, right?

Greg Lawton: 1:25:17

You exactly. So I just say the business case is just deploy it on two or three projects and have some incredibly set objectives around automation and transparency and these kinds of things. Have your scientific forms about, you know, this is what we're gonna measure, and this is how we're gonna measure it, and this is when we're gonna measure it. Do that incredibly because by the way, I say two or three. If it takes longer than two weeks, I can't be bothered anymore because we should have all of our answers within two weeks. And then it's like, okay, cool. And then you just you just go with us. So I'd say those two things, it's it's the um it's gonna really about the strategic objective of the organization and then rapid testing and execution.

Daniela Kellett: 1:26:00

And I think it's particularly easy with your platform because if you really think about what you have on offer, those things are things that are happening repetitively on a daily basis within your project, right? So it's very easy to measure apples with apples. But then the second benefit is these are things that we are potentially not doing on our project, but that the platform does do. And it gives you a whole new perspective on your project and how you can, you know, how you perform. So I think it's it's quite good. Um all right, so just one final question for you. As a platform owner in a project-driven world, are you still learning with every project or with every software iteration?

Greg Lawton: 1:26:43

Of course, of course. Like, I'm just again, we spoke about this at the start, we spoke about this beforehand. Like, this is all just a game. Like, the the only the only thing I really care about is is learning and getting to the next level of the game. And I define the next level as I've learnt enough to beat the final boss with the resources I have, so I get to start the next level at full health with more resources, but I have to tackle a bigger final boss at the end of the level. So, like that that's the whole game of this. Now, what is the final boss? It will be a problem. It'll be a business problem. It'll be like, well, how do we tackle more advanced claims? Or how do we tackle you know, these particular processes, or how do we tackle putting a wrapper around this particular AI to get ROI for the customer? Like it's just these things. So yeah, if if I'm if I'm not learning, then either I should be fired or I'm dead. Those are like the two things. Beautiful, there we go.

Daniela Kellett: 1:27:53

Listeners, if you're feeling inspired, check out the nodes and links website for a demo. Subscribe to Nodes and Links Academy on YouTube and the Beyond Deadlines podcast.

Daniela Kellett: 1:28:05

And that's a wrap for today's episode of Powering ProjeX. As we close out, I want to say one last huge thank you to Greg Lawton for joining us today. And thank you to all of our faithful listeners for tuning in to this episode of Powering ProjeX.

Daniela Kellett: 1:28:20

If you found value in today's episode, please do subscribe, share it with your community of project professionals, and leave me a review. Your engagement helps me reach more people just like you who may benefit from our community.

Daniela Kellett: 1:28:34

But don't stop there. Engage with us, share your challenges, your successes, and let me know which skills and techniques you'd like me to cover and help me discover the voices and stories that will benefit our entire community. If you have a suggestion, comment below or email me at poweringprojex with an x at gmail.com.

Daniela Kellett: 1:28:55

Until next time, keep powering your projects from vision to victory. Remembering that AI agents aren't here to replace you, they're here to partner with you and elevate you, turning complexity into clarity and delays into deliverables. So keep powering those projects towards victory. It's DK signing out. Bye.

Daniela Kellett: 1:29:17

The content provided in this podcast is for informational and educational purposes only, and does not constitute professional, legal, or financial advice. The opinions expressed by the host and guests are their own and do not necessarily reflect the views of any affiliated organizations or sponsors. While we strive to provide accurate and up-to-date information on project management methodologies and related topics, no guarantee is made regarding the completeness, accuracy, or applicability of the content to your specific circumstances. Listeners are encouraged to consult with qualified professionals before making decisions based on the information. Shared in this podcast. The hosts, guests, and producers are not liable for any losses, damages, or outcomes resulting from the use of this podcast's content. By listening to this podcast, you agree to assume all risks associated with the application of the information provided. Thank you for tuning in, and we hope you'll find value in our discussions.