Robert Harms

Hello, my name is Robert Harms. I’m the president of Lights in the Attic project group, or LitA for short. And we have a number of business units. One of those key business units is a company called Iscient and Iscient and focuses on the bridging the academics, the science, the technology, the bench science, that is just in the head just ahead of the the the work that’s being done are going to be done with respect primarily to artificial intelligence. And whatever that means. Today, what I’d like to do is deconstruct to demystify the the notion goes the term AI, and into translate it to break it down into certain two to four different definitions. So that we can we can talk about what those are, and where there is really application where we are working, and where we can be working and should be working across, particularly with a focus for capital projects. And the larger the project, the more fun it is in red across North America and the world. So I’ve got a number of slides in here that I’m going to talk through, and ultimately, ultimately leave with you. Just to to help everybody understand this. Or remember, hopefully, the term, the name of the company, Iscient and it is really supposed to contraction of two names. And the one, one word is is prescient. And so we’ve taken the suck this sient off of that and make it Iscient. net, it’s icy, I understand, and I’m looking forward. This specifically, what we’re going to in what we call interim AI, is the recognition that artificial intelligence is a very large and grandiose and even, you know, ambitious and an agreeable term, even with folks like Ray Kurtz will and and the folks who, in for the last 20 years, 30 years have really worked on the architecture of the concept and, and even this group, very esteemed group cannot agree on what AI is, and whether it’s going to take five years, 50 years, or 100 years, or even more to reach full AI, which is called the AGI artificial general intelligence. So what we do is we deconstruct that and say, you know, you know, the AI that is being the term that’s being used by marketers and people around the world to kind of inject some mysticism and some blackbox into saying something, you know, really cool is happening behind here, and we’re not going to tell you what it is. But so what we do is we break it down, say, Hey, we’re, what we’re looking at is a trajectory towards more complicated mathematics systems, tools, and the technology and the Applied Technology that work toward automation, attend autonomy to some degree, that are all still within the purview of an in the control of people. So we’re not going to pretend that there is an AI that is it’s out there and it’s spooky. And with the focus today, although there’s you know, this comes from a broader perspective, the focus today here is on capital projects, and pet projects, and he really does have a lot of implications, but for the larger company for engineering for engineering firms for constructors, OEMs, etc. But it really is this, this focus is on projects. So breaking this down, starting at the very, very beginning, what we call AI zero, and there is nothing artificial about this, this is a is for anthropic, it’s, it’s the, it’s the people and, and what we know what we do in how we in how we think and, and we how we apply that, just for the for a simple illustration here when we talk about AI zero.

If you remove the hand from this, this image, you’re left with another image and automatically we can see depending on the definition of your of your screen, that there’s a finger behind there, and because the hand is there, we automatically See that as we know, that’s a head and a body, presumably a man, just by the, by the the form of it. But even at that point we can, we can be wrong, but the computer and software won’t be able to, if you take the hand away, it won’t be able to know what’s back there, it could be a shadow of a bottle with a bottle cap, or a water bottle, etc. And, you know, I think we need to recognize, first of all that meaning is is fundamentally human, there is no meaning in artificial intelligence. Ai one AI A is for augmented This is when we’re collaborating with somebody, using a tool such as a, you know, books in the library, or any kind of collaborative tool communication tool where the information and intelligence that we use and get is, is not completely our role as an individual AI to this was where really work started in business in, in the early 80s. With and, with, with software, with the development of, you know, for example, Microsoft suite and the other things that were around at that time and, and designs, computer aided design software. So this is where it became automated as in download word sorry, software that was were executable files on our systems, which with a lot of us still have, and now into the cloud, where the the the tools are available in a much broader base and easier and creating mobility around the world. But that’s all automated. So automated takes instructions, it’s a hard wired, so to speak a set of instructions that that enable us to do things faster. And more productively we, we hope. Ai three is where it gets very, very interesting. And this is where we are starting. We are not in an artificial intelligence environment. And we’re not going to be in one for for some time. Artificial Intelligence here is when you develop a system, and yet it learns from the data patterns that you don’t plug in. So we learns patterns, everything comes down to patterns, because it doesn’t have any conceptual capability. It has only patterns of data. And it’s it’s how those get represented going in and coming out. And the the propagation systems back and forth, that understand that find patterns that can look chaotic to us. But but they’re but they’re deeper. Ai for is is a is for autonomous, and this is where we, this is where the Hollywood likes to, to spend its time. So it’s we’re not, we’re not even considering that at this point. So again, if we look at where we are AI zero, Ai, one collaborative and software, programming, coding, automation, and then artificial intelligence, and in the far distant future, so we’re here, we’re here. And we’re still playing around with here, I mean, in industry looking at the, you know, platforms, digitization, this is this is all in here, digital digitization is not in here, digitization is here, this is where we also are of course, and we’re just starting on this one. So what we break, when we look at introducing AI concepts or AI three concepts and elements from the rest for the rest of the conversation, I’ll just use the term AI, please recognize that I don’t mean it the same way as that the television means that you can get a smart map for your welcome mat for your for your door and pay and it’s built in with AI. So let’s let’s I will use AI in the way I just just defined it. So we use functional ladders we consider this is this is the thing breakthrough. This is just the way we think about it in that you take one step at a time but you’ve got your eye on the the the future moving further up. But it is not something where you can you can take AI and says say it’s going to solve all this.

It worked for today’s president What I’d like to talk about I put it in the in the sense of like project management portals. And this in this case, this is not a portal of not necessarily a portal, which is a screener way to get and this is, this is a way to see something, it’s a way to see things that information and insights that can help with our decision intelligence can help with making decisions, understanding risks, managing the business, that that we’re in, this is not a technical, you know, system in particular that you can go and buy. That’s, that’s not what’s intended here, this is a way to see the information that you need for project management. So we’ll talk about some here, some near horizon functional ladders. So these some of the things we’re going to we’re going to go through 10 items, and some of them are very near some of them are things that we are, we as industry are doing presently doing, some are a little louder than than others. And but a lot of these things are things that I wrote in 2019 wrote a paper about where we can go and where we should be going with this stuff as real practical applications to help us to leverage these, the potential and leverage these tools. So they vary in from current to the next two to three years out, I do not get into today will not get into quantum software, quantum computing, although we do certainly have our eye on that. And that is about a three to five year program that requires a separate discussion. So do keep it simple here for at least for me to talk about it, these things are all exciting. So I, I tend to get too deep into that. So what we’ll go through is each of these, these 10 items, relatively quickly, there’s a lot of words and meaning compressed in here. So I’ll go through that try to, to unpack it and, and communicate what I mean by those things. And recognizing that for some folks it’s going to be it’s not going to have an application and other folks it could be very, very exciting for for your business. So in in the EPC and project environment that one of the first things that that we can do with a with better information engineering, is rapid remodel of capex and OPEX consequences of alternative standards and plant redundancies and layouts that the tools that we have, you know, pretend to do this, but it is what we don’t have is, is the the pattern is zation that deep systems and the deep learning systems that can turn this into a quick models now it’s not all just about brute horsepower, it is about a different way to find those patterns and those consequences without having to go through an infinity of linear equations and, and, and an intervention by with engineering. It’s complex, and that’s why it is we haven’t been doing it yet. Number two fluid design model integration in iterative study with OEMs original equipment manufacturers and vendors for equipment and commodities. Now, what this means is is moving away from the traditions the traditional model where where your engineers, the operations engineers, as well as the project engineers begin working with, with different OEM some of them are, you know, preferred to equipment suppliers that the owner has, and others or others are going to be new, but the process of integrating the models and in and the design and working with more holistically, these are opportunities where and more dynamically. These are opportunities where where AI and some of those pattern recognitions and making that fluid where the it is doing a lot of the calculation and we’re increasing the OEM scope and responsibility in the overall plant design. And where were like, even from right down to two valves, two pumps to pressure vessels. Well, the people who who know those best are the OEMs.

And we tend to we as an industry tend to not really dismiss but we don’t really capture the full value of the OEMs whether they’re a large company like GE Or were there a small, a pump fabricator, in your state or in your province. And those, that’s where the people, that’s where the engineer the engineering and the knowledge, the intelligence expertise, and the data is for for your, for the, the project design. And so by getting a more fluid design model integration with that, we can not only be smarter about things, but we can be faster and we can have greater certainty, and in the operating teams for owners can, can be far more engaged than than we have been previously. Number three, to patronize and optimize engineering and fullcycle supply chain workflows. This is where the sister company for icy and partially I shouldn’t say sister company but another company we’re associated with the EPC lens is is right now has been working on and has developed tools for this, but it is we’re still in the front end of the the notion of pattern icing and optimizing the work cycle the workflows in EPC between an owner, the engineering firm, they the different consulting firms that support a central engineering firm, the OEMs the contractors, the fabricators, the manufacturers, that module yards, rig up all the way down to all those 1000s of points of, of input for with respect to materials and support for a project, you know, given any project at any time, a large project can can have five to 10,000, individual hand pairs of hands working on it around the world. So a lot of that, or the key parts of that can be can be optimized, and we’re still working in a in a stone age. And I don’t mean to be derogatory, it’s, you know, obviously, it’s an exaggeration, but it’s like we, we it and people get offended by now, I don’t mean to offend anybody, this is an opportunity for us to to optimize workflows. And part of the problem with that is that project people are always doing projects and it’s really difficult to, to rewire a company, an engineering firm and owner who is who were capital intensive. This is, you know, part of the difficulty with doing this so we’ve so EPC lens are working on that and it’s the internet the findings are quite interesting number for a live dynamic model of resources with business case analysis with environment specific, fast simulations, meaning that it is instead of looking at when we get into a project beginning of the project right through to further on when you want to take a look at making decisions around the resources that are available to changing conditions, particularly if your project is going on for more than a few months. If it’s going on, you know, as long as you know, to 3d even you know, even longer than many years the resources the environment that that we work in is completely different than the one we started with and the one we based all of our assumptions and on and the the strategy so having something that that works with the the changes in the in the resource environment. Data is with with a fast simulation that creates the opportunity to make much much smarter decisions and increase the confidence of predictability and success for the for the owner. Number five, test and prove out alternative organizational models for integrated owner, project management, organization engineering, construction and commissioning. Now, this doesn’t mean that the your model is wrong or however you’re doing this is wrong but it’s to test and prove out in an in a virtual environment in a calculated environment based on the patterns that we find with with workflows and how things get worked, including liabilities and responsibilities and how that can best be done. Between in the in the entire organization that’s going to deliver a capital capital project that can be a small company that does projects irregularly, or a large company that wants to look at it do a new model for how it’s going to deliver deliver projects.

Number six in real time market analysis into project controls and supply chain management out this is this is something I’ve been working on for some time in my background, 25 years building projects, this is a very, very painful start to trying to be right and almost always being wrong about decisions and understanding what the market is doing. And there are a number of most companies have things like category management in different processes for to try and to calibrate to measure and get information that they think that is going to give them an advantage and help with with decision making. And to some degree, that’s, it’s true, it’s extremely expensive, it’s slow. And it really doesn’t, there are a lot of lot of things that it can do better with it with a slightly different approach. And so we’ve done we’ve done this, we’re working with this now and rolling it out with the PC lens. But the power of this is quite remarkable. And it does change some change some things with with how a project can get done, which isn’t always making everybody happy. But it’s it’s a very, this is another strong potential for using engineer to algorithms and a different work process to to improve projects and and the OPEC side as well. Number seven, integrated risk modeling and contingency mitigation management systems, dynamic behavioral patterns and metrics, a lot of words, a lot of those words have, you know, a lot of baggage for us and, you know, looking at well risk modeling contingency mitigation management systems, it’s, that’s, uh, you know, it’s going to make people’s heads hurt, and mine included. But there is a real potential here to help. And we’re not we’re not working on this yet. And we don’t know anybody who’s doing this. Even of all the major, major technology companies, this is something where the certainty and predictability of projects is largely contingent on this, when you look at the in post project, post mortems of projects that have gone wrong, and there’s a lot of them. There’s a common, I guess, a common response by project teams and owners to say, well, it’s poor planning, so much stuff is poor planning. And I think we’ve got the question wrong to a large degree now, whether planning was, you know, at fault or not? That’s, that’s a whole big complex question. And we don’t have to actually ask, the question is, you know, how much planning Do you need, this is what we need to focus on here, what we can focus on now, with, with the technology and the mathematics and engineering these algorithms is to actually be much more aware and plan for the things that are changing. And the things that can change and the things that you may want to change, to be able to look at those, all those, those dynamic elements in a more real time basis. And with that, in as the future goes on here that will get you know, the time required to do that is shorter and shorter, but at least to have a dynamic system between what’s happening. And that is calculating, you know, working with your business objectives and your risk model, you have a power to take decisions before something goes awry, before something goes sideways or pear shaped. So that having that decision information is is critical for bringing project costs way down and getting even getting rid of some unnecessary contingency and being able to manage that much much better with greater predictability and to be able to go to our board of directors or our senior executive and say, you know, we we do believe we know what’s coming and we do believe we have a we’re ready for it. And we have the tools in order to to manage it.

Number eight dynamic live resource benchmarking multiple metrics. And this is this is having an ability to to calibrate in a with a greater confidence and accuracy using a number of different key metrics and KPIs. Not only benchmarking, you know, there was productivity and costs. There’s there are a number of things We can do and I know there’s a lot of art, there are a lot of arguments, and I face those arguments, you know, in the business, and how do we approach things like, well, benchmarking requires a universal definition that is simply impossible to achieve. And, again, it is not easy. And, and it’s generally left at a very high level, which often doesn’t give us a whole lot of information, actionable intelligence in order to, to make changes or understand where we are. So this is something that I hope that that we can, you know, move into. And certainly industry does need to do this. And the more that everybody gets involved in that, the more the better it is for everybody. And the focus can be not on competition so much between businesses, but actually just bring your own absolute value down to as your your absolute values as high as possible. Number nine, live capex and OPEX projects, market intelligence. And so again, this is something we’re supporting EPC lens with, and their tool called the tool called marketscape. And that is just understanding what’s going on not on the high level? Well, it is on a high level, it’s scalar. It is at a micro level, intermediate level, regional level, pardon me, national, state, provincial and international level of what’s going on with, with projects and the cost, the trends, the availability of resources, cost inputs, different kinds of elements on risk, and sentiment, the people who, you know, market analysts tend to think that they’ve got everything in hand in the they put out a forecast, and then the then reality transpires and, and in the end, we’re left with some, some rationalization and some excuses and some things that were unforeseen, that explained the differences between the original forecast and, and reality. And with the, and that’s great, and there’s some terrific intelligence tools and and people that do that. But what what this tool does in this process, and where it’s working with the AI, the algorithm engineering is, is getting data from the ground level from the grassroots on a continuous basis. And looking at that, and looking at the pattern so that the reports, whether it’s just basic data, or it’s more comprehensive information, or it is true analysis, you know, based on on complex patterns that we normally otherwise wouldn’t see, and in correlations to data that just simply are too complex for for market analysts, to conduct, this is the stuff that we can have on a daily daily basis. And on portals and dashboards, so this is obviously something that is has been silently in progress for several years and is very exciting. And from that, a number of these other tools become possible, again, to be looking at that, that that ladder process, you get a few of them at the at the bottom level, and it creates enormous possibilities. And you can decide your your own trajectory with depending on your own corporate objectives. Lastly, number 10. The Iterative iterative pattern recognition of the lessons learned programs and collaboration systems now, there’s most of us feel the pain or have felt the pain of, of not only trying to capture lessons and things that have happened, whether they’re smaller, or or catastrophic.

There’s a whole theme of that meant learning management process is difficult as it is, but to get it right and to to not blow things up a second time is you know, that’s where the value is in. We’re struggling with that. Not only with the with the industries at large, but even within an individual company, is just how do we make sure there’s that continuous education that recycle or the feedback loop with lessons and this is one place where you know this, there’s a terrific opportunity here for the technology and the mathematics and the algorithms that are already available. That toolkit can be used to do something much, much better here than anything that I’ve seen to date. And I’m assuming that there might be some some silent work there behind the behind the curtain but it is, this is a terrific opportunity for for all organizations, including the engineering firms, in construction firms. Other things, this is something that’s a little bit more on a strategic level. And that goes back to a statement made by the CEO of IBM A number of years ago with their strategy on how they’re going to design, the what the architecture of their solutions are moving forward, and they really avoid the term AI, particularly high level conferences. They use, otherwise, that add more consumer level. But what their approach is to go is, is what I mentioned in the AI two is, it’s really augmented intelligence. And the way they want to do this, or the way they are doing this is, is looking at all the different functions as microservices, and then creating a platform that looks at all of those different things. And this is you can kind of see this with API systems, etc. But it’s a platform that integrates those things. And that can, cannot, can make them work together holistically, which is, which is very, very interesting. And it makes sense. But one of the things that we can do that for organizations, specifically, you know, stepping away from the IBM, some large ones is it’s it’s the, the algorithm based, deep learning form of having a consultant coming in, looking at a work process, working out looking at a workflow, whether it’s supply chain, project controls, how you’re managing finances for project, etc. And while you’re continuing work, you have a you have a team task force or consultant come in and look at that. And then they say, well, these are things, these are areas where you can improve. Well, you know, and that’s great. And there is, but it’s, our operations are increasingly complicated, increasingly large, and increasingly, the pace of them is, and change pace of operation, the pace of changes is so quick. And that how do you how do you learn something about what you can? Where you want to make changes? Where you how do you get to understand the ROI, the nature of that change? how much it’s going to disrupt or impact your organization? And, and, and what’s going to be the return on this is what’s going to be the, you know, the end game on that? What’s it going to look like? Well, this is something we’re deep learning parallel, deep learning through these, looking at each different modules of operation can really, really help Quicken that accelerate the pace of that find things and actually evaluate them, including the consequences with much, much greater accuracy, then we can press something the way so those are the 10 items, that they’re not comprehensive, they’re not exhaustive by any means. Or there are, you know, it’s it’s a barrelful official fair. And so 10 items, and and the notion of how we can move forward with some others, there are a lot of other opportunities, but generally what the way we need to work this now it’s, it is a vertical process, it’s a vertical and a horizontal process. Whereas we, typically since since the early 80s, and the we saw the need industry adopted the the, the, the model of having an information technology or Information Systems stratum, across the organization’s to support software tools, the hardware, etc, even communications, tools, etc. Well, what I, the way I see this is going to happen is we’re going to be adding another stratum here that is horizontal and it it is a it’s really algorithm engineering for business intelligence and decision based.

Intelligence based decision making. And, and so it’s gonna it’s different people, it’s a different skill set. It’s a beast. It’s not the term data science as it’s being used now, but it is it is really truly science and data engineering, algorithm engineering, combined overlapped with, with business management. It is it’s not so much how the microwave works. It’s what you put in it to cook. So we do this with theirs. It’s a top down approach, really, it is starting with the corporate vision, the research, and then it goes into the theory, the things that are happening in our universities today. And how that turns into Applied Science and then into technology, using the ladder concept to go one bit at a time, but having a, a long term trajectory. And then looking at the adjacent elements, the factors because as we become more integrated, and the data becomes horizontal, excuse me, we have to be much more much better at looking beyond a single business unit looking beyond a single discipline. And having a being mindful, I guess, of the the adjacent impacts and adjacent industries, like there are things happening in medicine and manufacturing that have immediate direct application to the things that we’re doing. And there’s some really terrific ideas there. So then how that applies to the solution designs, we can finally do the solution design engineer, and actually get into tangible, real application. So those are vertical concepts, and there’s a lot to be done there. So it’s you have to take all these different strata, or strata, and and look at it that way. This is no longer a, we no longer tell the the IT group to go buy some software. So thank you for the time for letting me talk about those things. I think that there is it’s a very exciting time, it’s a very anxious time for a lot of companies. And perhaps some of the, I think that the fundamental notion here is that it is while it may seem daunting, there are things that we can do that, that that have a terrific return, that won’t send us sideways, you know, whether it’s a company or an industry, and we just need to recognize break this thing down into, you know, what will, what do we what can we actually do? And in how can we do that? So today’s discussion comes out of a, an education series that, that that I’m working on, to deliver it’s called interim AI based on a white paper, an E book, actually, I wrote in May 2019. And it’s available. Pardon me. But the other some other things here that you can see that we’ve been doing we have been working on particularly in Canada is the indigenous lens here, which is a applying a lot of this stuff to understanding the the Aboriginal indigenous, First Nations, Native Americans, etc. How projects can integrate with with these communities. And a couple of the other things that that I’ve done there, one of the things I’m quite excited about, as well as what’s coming up is reintroducing a group, a nonprofit called capex science. So I’m hoping that I get to talk to you about that as well in the future. Just a little bit about ICS where we’re working you know, basically it’s the engine for EPC lens. My group leader PG lion tamer which does the execution planning and the turtle, which is a new development with an indigo indigenous group in Canada called Turtle Island Workforce Solutions and how they’re creating a fabulous organization of trades across Canada with a using our systems to to create visibility and management systems.