Meet Nodes & Links: The AI for Project Controls

00:01 Hello everyone, what is Nodes and Links? Well, in one sentence, it’s about making AI in project controls a reality. I’ll dive in, and you just sit back and enjoy. For the next twenty minutes? Not twenty minutes for this one presentation. It’s broken down into a couple, take it at your own pace.

00:33 So, a little bit about me. I’m Greg Lawton, I’m the CEO, I’m one of the two founders of Nodes and Links. My background is in physics, then into engineering and project management on the UK’s nuclear submarine fleet with BA Systems, then into venture capital and technology building. That is where I founded Nodes and Links. Nodes and Links is AI-powered project controls, and in the next video, I’ll explain why that became the mission of the entire company.

01:19 I say to my teams, we just need to do two things. We need to save you time, and we need to make it very easy to do advanced analysis. And the benchmark that I hold my teams to is at least ten hours per person per month of time saved. That is irrelevant of the acceleration, irrelevant of the risk reduction, irrelevant of the benefits of better stakeholder engagement, irrelevant of the reduction in risk of going to quarter, the increase in the probability of winning the claim. All of that is irrelevant. It is very just at the heart, we have to save ten hours a month per person before we start doing all of this.

01:59 We sit on top of your scheduling systems, so you can have P6, Asta Power Project, or Microsoft Project, and we take in those files, and the system automates the analysis of them for you. That is why saving time is at the heart of everything. It involves a huge swathe of the different analysis that you need to do: health, progress, productivity, resources, trending, path comparison, delay, risk predictions, acceleration, and carbon. And more are coming out every single week. The ones that specifically have very advanced AI associated with them are risk predictions, acceleration of carbon, and of course, the AI assistant and your AI reporting, which both together make your copartner. Your copartner can guide you where to go in the system to do the analysis you want, and it can do analysis for you at your instruction.

02:37 We’re used across the world, so at the moment of this recording, it’s in twenty-three countries with contractors, consultants, and clients. We’re growing at around three to four hundred percent a year, and I’m thrilled that you’re part of this journey with us.

00:01 So the founding story of Nodes and Links is actually really quite simple. Myself and my co-founder, a gentleman named Dr. Christos Ailamakis, who is a researcher at Bristol University in predicting projects and a visiting professor at Stevens in the US, we were sitting down and we were talking about the future and the fate of projects. And two things shone out to me. Number one was a recent report from the Global Infrastructure Outlook Initiative, which essentially says that the amount of money that we need to invest in infrastructure to maintain
00:39 the same quality of life given the increasing demand on the climate and the environment and the growing population of humanity, the need is far in excess of what we actually have. So, we need more infrastructure than we currently have and then we’ve currently got planned to maintain the quality of life and to protect the planet. The second one was about delivering these infrastructure, which was reports from across the world but here’s one from the Association of Project Management, which essentially says we don’t have enough people. It’s
01:17 saying on an annualized basis we need another 2.2 million project managers a year just to fulfill the current demand. So what we’re seeing was a future where our quality of life decreased on average or certain patches in the world just were left us to decay whilst the first world continued to prosper and on the other, a lack of actual people to deliver just that basic case and we thought, well, we need to change this. What we need is we need a lot more capacity in the system. We need to dramatically increase the productivity
01:59 of humans that we can do more without more stress and without more effort. We need scale technologies. What we did then was have a look at how project controls has evolved over the course of the last ten years. Put very simply, when I started learning project project controls looked roughly something like this. So you had scheduling cost risk change and controls within leadership project managers controls fed leadership with information project managers provided controls with information leadership put down directives, etcetera, etcetera.
02:41 Now today in 2023, this has changed and it’s changed in one key way: the top controls teams around the world have employed data scientists. Now, what this actually means is they started putting their data into databases and starting to analyze it and almost every consultancy is providing some kind of dashboard. But coming back to the vastly increasing the productivity of people, we thought we need more than people in the workforce. We need a transformation that’s happened elsewhere in the world with machinery but we need
03:22 it in the office. We need every single one of these people to have an AI system. We can vastly increase the productivity of people by giving people workforces to work for them and that was the goal that we set out to do. We thought by vastly increasing the workforce that was working for humans we could make up that gap. It would lower the cost of delivering projects so we could get closer to the investment trend that we needed. Now we were the only ones thinking this. There’s a number of research articles especially in 2023 that’s come
04:07 out essentially saying that the future of project management is artificial intelligence in some way. Now we have some very, very strong morals and beliefs on how the future of project management becomes AI which you can watch in a video later in this series but the fundamental thing is this is that it is absolutely inevitable and it’s inevitable because of the demand there is in the world for more infrastructure. People aren’t going to accept not having running water and accept not having electricity because of standing on principle
04:49 about their beliefs in AI they’re going to demand those things with moral standing alongside AI. So with all of that we set out half a decade ago to do one simple thing. I said to my teams we must create the AI system that works with or even better for people to deliver the world’s projects. That is Nodes and Links.

00:01 So what is artificial intelligence for in project controls? Well, to answer that question, we’re going to have to go through a story, and it is the story of how technologies evolved since the dawn of humanity. Now before we start, I need to ask you a question: How is a fishing net an artificial intelligence the same thing? Now, I advise that you pause this video and I have at least two ideas of these because it will be enjoyable and we’ll end up learning something. So let me get into this. Let’s have a look at how technologies have worked since.

00:39 The dawn of humanity. Every single piece of technology that humanity has ever invented can be described via a contrarian question. So I want you to go back and imagine that you and I are building the pyramids in Egypt five thousand years ago, and I’m saying to you, building this pyramid is a hard thing. I have to remember so much stuff and I forget it. I’ve tried to make diagrams in the sand, but the wind blows it away and people walk over it, and I don’t quite know how to improve it. And then you say to me, “Well, what if I could… What if I could tell you that you could save your thoughts for people for five thousand years in the future to understand and give instructions to people hundreds of miles away without any misinterpretation?” I’d say, “Well, that’s crazy,” and you would introduce me to the Papyrus paper. To paper. Fast forward to the internet. What if we could send and communicate instantly between people and machines across the world? And now to artificial intelligence. What if I didn’t have to do all of the thinking? Now we could go even further and say,

02:11 Well, having machines automatically create assets for us. What if we didn’t have to construct the walls ourselves? We’re getting into robotics there, but you see the point. All of these can be described as a contrarian question that people would have asked before the technology became commonplace, and all of these are progression of automation. So let’s dive down deeper into that. What is automation? Well, automation is about scaling the productivity of humanity. With a fishing net, I can catch more fish in an hour than you can with your bare hands. With an engine in a car, I can travel further and with heavier loads than you can on your feet carrying things on your back. With artificial intelligence, I can create more thought work in an hour than you can with a pen or a keyboard. You take a human and you scale the output of that human with technology, and every single piece of technology you have in your home life or your work life does this. So how are these two the same? Well, they scale the productivity of humanity. They scale the productivity of me or you.

03:37 It’s just that they scale it in different ways. A fishing net scales muscle. AI scales mind. In the past, the industrial revolution has given us a number of technologies for scaling muscle. We wouldn’t think of carrying all the concrete from the concrete factory to a construction site without using a truck. That is standard practice. What if I didn’t have to carry the concrete? But we’re at the dawn now of having automation of the mind. What if I didn’t have to do all the analysis? What if I didn’t have to do all the planning? What if I didn’t have to write all of the reports? That is the era that we’ve just entered. Now, what does this do? Well, all technology fundamentally does more things on one side of a graph. You’ve got people and you’ve got more or less are the same, and on the other side, you’ve got work. You’ve got more or less or the same or better or worse, etc. Now, what technology does is we can take the same people and produce better work. So think here, there’s a software program called GitHub, and Microsoft

04:51 Released GitHub Copilot a couple of years ago where an individual could just say, “I want code that does this,” and it would draft the first line of code, then a person would go in and check it. That was significantly more productive than people writing it from scratch. We’ve got the same people more work, which is the concrete example and the fishing net example. We’ve got more people less expertise. This is Uber before Uber. To be a taxi driver in London, you had to pass the taxi drivers test. Now all you have to

06:06 Do is download an app and Google Maps tells you where to go, and we have less people for the same work. This is farming. So if we go back three hundred years ago, the vast number of people in the US and UK populations were farmers. Now it’s less than five percent. Every single change just leads to better unit economics for every pound or dollar we spend in terms of people’s time or technology, we get more pounds or dollars out after the technology than we do before, which comes down to the fundamental reason of why is this

06:49 Absolutely inevitable in project controls because of money. The average salary in the UK US and Australia I’ve got UK and Australia here if you can get significantly more work for a slight increase in pay which you will have because of the increased skill setting in planning and controls you will do that as a company and because like in the last video we’re two point two million project managers a year down on overall demand there is more demand than there is supply so supply can flex to equal demand that is why artificial

06:49 Intelligence and project controls is absolutely inevitable.

00:01 How do we think about AI design at nodes and links? Now I’m not talking here about how we think about coding in terms of X comes before Y or Y before X. What I’m thinking about here is how do we actually think about AI design for you? How do we decide what to build and what not to build? And not is the most crucial question, and the answer to that come with an understanding of what a role is. So this is the framework pretty much of every single role in humanity. Now I’ve just taken this as an example. It’s a raw profile of the evolution

00:42 Of a scheduler from a train a graduate or a trainee to an individual contributor to a manager to a strategic leader where you’d be looking at things like policy across the organization training and hiring etcetera and what characterizes every single role that I’ve ever seen is that every role is a collection of responsibilities. You hire someone to be responsible for something. Responsibilities are delivered via workflows so for example if you have the responsibility for monthly reporting you will have a workflow where you collect data

01:19 Analyze data create report have it checked and verified send report out maybe give a presentation that is a workflow and workflows are comprised of tasks and it is at the task level that technology automates this is not artificial intelligence is just another technology it is a new tool that we have before we had lots and lots of tools that could automate our muscles now we have tools that can automate the mind and yes things like keyboards etc are actually partially ways of automating muscle because they automate you having

02:01 To write to you having to type etc now if something’s typing for you that is automation of the mind now what we do here is look at all of the workflows all of the tasks and we figure out where it is beneficial for artificial intelligence to come in and beneficial to you how does it save you time give you a better answer or make you look better those are essentially the three criteria and I’ll give an example of this so a quantitative schedule risk analysis some of you might not be familiar with this but it’s generally the process of how you do

02:40 A risk analysis on a schedule the process is like this you prep the schedule you’ve got a large schedule fifty thousand line items a planner would prep a smaller schedule something like five hundred line items people would then manually assign the inputs of the analysis which is uncertainty so they’d say duration wobbles and risk events they’d then run a simulation they’d then take that output and decide what to do with it so commercial can take the output and decide what to contract with it and risk

03:14 Management and project managers could take the output and decide how to de-risk the project or accelerate it and then that process would be repeated in a periodic sequence with artificial intelligence we save time so we don’t need to summarize fifty thousand line schedules because AI can deal with volume of data very easily it saves us time using past data and machine learning we can recommend uncertainty distributions and risk events so that we’re getting we’re automating elements of that point in the process but

03:50 We’re not saying what they exactly are because part of this process is that there’s a workshop held and people gain agreement and understanding that is more important match with the underlying answer there’s then automations around reporting and writing of the findings and because that entire process which can take hundreds of hours can now take ten twenty hours it can be repeated much more frequently to the benefit of the project AI automate tasks you bring tasks into workflows and design workflows

04:27 To deliver responsibilities and there are many responsibilities and it is that exact fact why AI can never do a project controller’s job or a project scheduler or a cost manager or a risk manager it’s because there are activities that actually are required to exist in the real world social interaction for example and this is part of that system that you’ll now start to look at when you go through the courses in action this for a twenty day duration activity the AI is suggesting an actual duration

05:00 Of twenty two days but I can turn it off and I can tell it to shut up if I want to you are in control you are in charge AI serves us as humans it does not dictate to us and that is a core principle of nodes and links and we can come onto some of the core morals and beliefs that we have in a moment