Hello listeners. This is the risky planner podcast. Thanks for tuning in. Hi everybody. Welcome to the risky planner Podcast. I'm Nate. I'm Albert. Hey, Albert. Hey Nate. Happy New Year to you, buddy. Yeah. You know I was thinking, I mean, I love the holidays like, you know, spending family, relaxing. But from a work perspective, no matter how much I prepare before the holidays, I always feel like I'm behind the eight ball coming back. Yeah, hundreds a year and more I talk to people about this and reveal my insecurities, the more I realize that I'm just normal and not Yeah, you know the exception. You know, what's fun, the flip side of that insecurity is my my personal experience with the holidays, where, when you come back and you're dealing with all this catch up stuff, and you realize that everybody else is dealing with all that catch up stuff, too. And then you think back to the prior three weeks, when everything more or less progressed as normal, and you start to wonder, what am I doing with my life? Yeah, if the world can just forget about me for three weeks and everything's fine, little little darker than I meant, but that would be, that's, that's the dream. There you go. That's the dream. Oh, you're right. That's, you know, I guess if you take that phenomenon and extend it out indefinitely, that that is called retirement, yeah, do you do anything cool over the holidays, visited some family in Seattle, handled some minor drama on that front as I'm sure many of you out there in the podcast listening world did as well, family family drama. I did not have family drama, thankfully, but I did complete a rubik cube for the first time in my life. I decided that I would solve one, and so I started memorizing the algorithms and nice. And then I also, you know, played around with trying to build AI agents. I mean, I'm in marketing, so I kind of view AI as sort of a real, like paradigm shift in my career. So before, you know, we used to, you know, you look at your career as in, you know, this is, this is what I know. This is what I do. And now with AI, it's kind of a force multiplier where I'm not just a kind of single T shaped professional. Now I'm like a multi T shaped professional, because AI allows me to be, you know, whatever I was, like 50% decent at, to be, like 80 and 90% pretty good at. And so there's a lot of enhancements that I'm finding. And so that's kind of a, kind of a lead into the show here. But you know, I spent time just playing around with with AI tools over the holidays, and it reminds me of a story like, you know, here in Toronto, there's this, there's this robot sushi restaurant. And so you go to this robot sushi restaurant, oh yeah, it's exactly how you would imagine. So you order, you know, via iPad, and then a robot with like, multiple shelves, kind of in its body comes out and then, like, its arms, sort of hand you the tray with the sushi that you ordered, and it, like, it's as gimmicky as you can possibly imagine. And there's, yeah, always, there's always a waiter that comes with the robot, just to make sure the robot does things right. You know, that sounds a little bit like when we talked about mining equipment many moons ago. We for you out there listening. It won't have been many moons. It will have been one moon, perhaps. But anyway, we talked about the need for humans to back up the automated systems and how long into the future that needs to go before, you know, people feel comfortable letting letting go of the controls and letting the machines take over. I'm actually, in a way, kind of relieved to hear that all the sushi robot does is deliver sushi, because as soon as you said robot sushi restaurant, I'm like, That's ground zero for the apocalypse. That's what that is. Because you we've just given extremely sharp knives to the robots, yeah, so they'll talk to you. I think, like my kids got a huge kick out of it, because it would talk, you know, not that like it wouldn't talk to you, like, it didn't have any kind of a enhanced sort of language detection, or, you know, it wouldn't, like, think of things to say, but, yeah, I would pretend to talk to you. And so my son, who was, I think, five at the time, just like, loved it. Loved, yeah, yeah, you know, because he thought it was talking to him. Right? So, yeah, it was, and you know what? In a way it was because most, most interactions with strangers can be boiled down to, you know, Stranger makes noises that you are meant to understand. But whether they, you know, cross your brain threshold or not is kind of unimportant, yep, yep. But I went to a restaurant as a soup soup restaurant. Primarily it was like, you know, Vietnamese soups. We get lots of faux and that's and they had a robot delivering their food. And let me tell you, robots are not up to the task of delivering a full bowl of soup just yet. Okay, much of that soup ended up just where you think it went, the floor. So did not seem safe or like good value, but hey, whatever. And also, we had to get the food out of the robot ourselves. Yeah, so the robot would show up to the table and be like, Here's your food. Yeah, you know, do as you will, yeah, you know. And just to go back so, so I was spending time over the holidays. You know, playing around with AI, building AI agents, following influencers on LinkedIn. You know, whatever you want to call them to try and develop, you know, my own agents to do marketing work. You know, content creation automation. But there's a lot of impacts that these type of tools and approaches, not necessarily apps, and maybe we'll talk about that too. Are going to have on like our work. So project management, project controls, risk management this year, cetera. There's a lot, and it's gonna, it's gonna start to really take off this year, I think, because last year, in many ways, was the year when everybody got comfortable with the idea that AI is a thing that they need to learn how to use and interact with on a daily basis. This is the year when we're going to start actually using it, I think. And like, I don't know about, yeah, actually, I do know about you, like, both of us, have been using AI pretty actively in 2024 for all kinds of different things. But it boils down to, like, utilizing a large language model to do large language model stuff. You know, AI is a vast constellation of interrelated topics and concepts and tools and things, of which we are only really hitting the tip of the iceberg. And to your point, like, it hasn't really been commodified and, like democratized yet, where there's still like a handful of tent pole options, all of which are large language models that you know, we interact with on, in some cases, a daily basis. Like you. I spent some fun time with AI buddies over the break as well. I was trying to see if I could get it to build like a Gantt chart visualizer, and I used Gemini, and which used to be known as Bard, I used chatgpt, I used copilot, I used Claude, just seeing which one would like generate Python code The best. And it turns out the answer was, actually, I got chat GPT to generate the code and Gemini to check it and add features to it. So, you know, there's, there's this whole and, like, I think that speaks to the difficulty of, let me, let me jump forward a little bit. Yeah, people are the things that are making the news right now are, like, jobs being lost to AI, and like CEOs saying they're gonna replace half their workforce, you know, with AI. Or, like people in the coding space, like, you know, people who are web developers or application developers, like wringing their hands about AI taking their jobs. I know a handful of web developers, very, very close friends of mine, they use AI as part of their jobs every day. But they, you know, you used a really good term for it earlier, as a force multiplier for them. Okay, it's not in any danger of replacing the job of somebody who's actually good at it. You know what? I mean? Yeah, if there's one thing, AI isn't right now, it's not a soup to nuts solution for any of these problems. No, it's not, and I think it's not necessarily. People keep saying, oh, AIs are coming for your job. No, no, the person who's good at using AI is coming for your job, absolutely. And that person is at 100% right now, coming for you, the listener's job. Yeah. So it's true. That's That's why I said, you know 2025, is the year when we all start using it. But, you know, you mentioned earlier, like, a gap in application space where there's, there's, like, conceptually, AI is out there. We can all use it, but it's in like, that startup mode, right? It's the same like when we all started using Uber, and we were like, it's so cheap and the drivers are getting paid super well. How is this making any money? And the answer is it wasn't right. And as soon as suddenly, Uber had to start making money. They had a really difficult time converting that great end user experience, you know, into an actual commodity that could make, you know, and keep making money. But they did do it. And. Essentially, it just took, it was a rocky road, and I think we're going to see that rocky road for AI pretty soon. Here. Do you think the prices on some of the maybe this is a tangent, the prices are going to go I guess they are based on experience for these types of services. Yeah, the prices will go up as people. I mean, they have to, yeah, because the vast majority of them are free right now. And you know, if you're if you're a developer and like your your tool requires a vast data set to draw from, to be trained and tested, then releasing your tool for free in its kind of nascent form is great for you, and it's great for me as an end user, because I get this incredibly powerful helper robot for free, CAPTCHAs, CAPTCHAs, right? Like, you know, training AIS for years, and, yeah, it was all based on cats, right? Like, that's where it I think was it Google that started, you know, CAPTCHAs. I think so, yeah, or not even that before that was really just training an AI to recognize cat photos. And so a lot of the technologies has now morphed into, like, you know, surveillance tech or facial recognition. So there's, there's lots of applications there. I mean, I so just in terms of AI tools, I think I have two, three, at least three subscriptions now to different platform like Claude. My favorite use case for Claude is content, and I use it as a personal assistant, right? Like, yeah, I have a project where I just, I do a brain dump in the morning, and I just sort it, you know, I have this specific instructions, and, you know, it helps me get organized, and then it actually builds me a roadmap. You know, here's different tasks that need to be done and amount of time that these tasks. It's an ADHD person's dream, right? But, you know, and then, and then Gemini, like their deep research, like they, killed it in December. They came out with a deep research model. They came out with a bunch of other models, right? And, yeah, complexity is killing it with, you know, the search AI enabled searches is awesome, and I'm using I go to perplexity on my phone. That's my first stop for searches. Now, you know, it's just so good, it's only gonna get better. It is. So let's pivot and talk about our space and talk about our listeners. Because yes, we did say that the people using AI are coming for their jobs. Don't, you know, don't, don't fret just yet. But there are some pretty cool stuff that like you're reading, yeah, for the project management space. So I'm seeing two things. Okay, one is a big old raft to companies that are claiming that their thing is an end to end, AI enabled project management solution like it's just a massive list of companies that personally, because I'm not an agile project manager, generally speaking, I do usually like traditional critical path method and hybrid project management. So I have to like, I have to know enough about Agile to be dangerous, but I'm not like a software developer, okay, so, but people who are software developers are using tools like Jira, Rike, Monday, Asana. On a more basic list, there's like task aid and Trello and those kinds of things like those. Those tools are all getting AI components, all of them, yeah, and a lot of, I would say the vast majority of the AI functionality that's already extant in those tools is about writing, summarizing and editing text. Okay, so it's the same thing that we were talking about earlier, when we weren't focused on project management solutions. Frankly, the large language model. When people think of AI right now, they're thinking of large language models, okay, but when I think of AI as a potential solution for my kinds of problems, like as a project controls professional it's all about taking historical data and analyzing it and extrapolating it into forecasting, you know, picking up on trends, identifying risks, those kinds of things, and they're passively okay at that right now. But here's the thing, the best stop for doing those kinds of tools, or, sorry, those kinds of use cases with these existing tools is actually the generic, large language models themselves. So there's still not been any real, you know, break in, into the, you know, Project controls, application space by AI just yet, at least not that I've seen. I mean, there's a few things that'll do, like, there's a software that's just called forecast, and forecast actually has what appears to be a pretty good use case for AI as a forecasting helper bot, as I said earlier, base camp will do like automated reports based on project data. There's also some middleware stuff that looks like it would be. Helpful for Project controls, professional like, step size will take all of your different project management platforms. It's focused on software, but, like, it's a good idea where, like, they're going to to get and slack and Microsoft and all those different places that have your tasks, putting them all into one place, and then pushing status updates back out. Smart stuff, okay, yeah. One Cal for taking calendars across platforms and synchronizing we work with if one Cal were like, totally secure end to end, it's probably something we would use, maybe not that specifically, but something like it, because we have clients that have secure systems that we and our systems are have a different security level to them, so we don't typically like synchronize calendar events across those two systems for security reasons. Can you talk about that a little? Actually, I have two questions. So that just a security, security is a big deal if you're giving it, if you're given, you know, that's, that's something that everybody asks and, you know, you have to read the fine. Can you talk about the security? This is a little bit of a tangent, I think, but an important, oh, it's not, though it's, it's, it's not a tangent, yeah, because, just like we were saying earlier, the the the way that these, these models are being trained right now is by interacting with their users, of which there are many, many millions. Because the tools are, generally speaking, free. Like all the tools I mentioned earlier when I was like, building my little Gant chart generator. They all have free versions. All of them do okay, part of the reason why I was using Gemini to check chat gpts work is because chat GPT is free implementation has, like, a limited number of responses you can get out of it, but like, the one thing that they all have in common, with a handful of exceptions, is that they are all using your data to train the models, and they're doing it out of data centers that are scattered across the world. So here in Canada, where we do most of our work, we work in Canada's nuclear industry, which is beholden to the Canadian Nuclear Secrets Act, which requires that any information that might contain Canadian nuclear secrets has to be housed within Canada. Obviously, the Canadian Nuclear Secrets Act was written in a bygone age when people were keeping their files primarily in physical paper form, in file folders and cabinets and stuff. And so those papers and folders and cabinets had to be in Canada, right? Well, the modern, 21st century version of that is that the data centers need to be in Canada. So like, the hard drive, or the tape drive, or whatever is storing your data needs to be here. Yeah, yeah. So that really complicates things when you are trying to use, you know, an AI chat bot that doesn't necessarily have, it's, it's like hardware stack fully in Canada. Yeah, okay, yeah. Like even Microsoft and copilot. Copilot is one of the most secure ones of these, because it basically is using Azure Security. I'm not a web developer or, like, a, you know, I'm not an expert in any of this. But, like, we actually worked pretty extensively with our IT professional here at occasion, company working on like, ensuring that the security levels that applied to our Azure cloud also applied to COPPA, and they do, yeah, so like, when you're making decisions about which AI buddy to use, security is a huge consider, and if it isn't a consideration for you, it's time to make it one, because you're If you don't figure it out right now, your clients certainly will. Yeah, okay, yep. So let me ask you a question about the so you actually gave a use case, presented a use case for just project controls and risks. So you talked about taking historical data, analyzing, extrapolating into forecasting. Can you put that into, like, a day to day, like, give me an example of, say, how a capital project might get value from, you know, one of these tools. Yeah, you bet. So, first off, it's worth noting that, you know, as I mentioned earlier, AI is a lot deeper than a large language model like AI is, is any, you know, and again, I'm not an expert, but like, it's anything that uses a gigantic pile of data and what's known as an inference engine to create an output based on those data. I just want to make a note here, Albert, you've said I am not an expert at least twice on this podcast. So you heard it here first. Folks, you heard it here first. If you're looking for a podcast that contains no expertise whatsoever, expertise, but Albert is truly an expert here. So let's Yeah, so let's talk about this. Use kids. Yeah, sure. Sorry. So one thing I am an expert in is forecasting methodologies. Okay, so when talking about how AI can help you with that, the answer is that, if you've ever looked at the gigantic spreadsheet of data for a large capital project before, it's a lot of numbers to crunch. Computers crunch them way, way better than humans do. Obviously, back in my Dad's Day, would whip out a slide rule and a broad sheet and do this all by hand, and we don't have to do. That anymore, thanks to excel and all of the many project management and cost management tools that exist, they'll do a lot of this work automatically, but you generally have to pick, you know, as the end user, how you want your project to be forecast. And a lot of the forecasts that come out of even the best cost control systems are pretty bad. So what I'm talking about when I talk about a forecast is like, starting today, your project has history that goes into the past, starting at its inception, and it has forecasts that need to go into the future for, effectively, the total life cycle of the project, you know, all the way through handover into operations and then eventually decommissioning of the asset. So if you're forecasting, you need to know a, what your time horizon is like, how far out you're trying to forecast. B, what purpose This forecast is meant to serve. Like, if you're just forecasting capital spend, then you pretty much take it out to hand over and stop, okay, and see what method you're going to use to forecast things. Okay? Because if you are, you know, using one of the even fanciest cost management tools out there, it's going to give you some options. Like, do you want to do a regression analysis on your existing data and then forecast your spend based on that? Like, do you want to make it multi like, polynomial, how many polys Do you want? Like, is it going to be, you know, a square or, you know, the power three, or whatever? Like, all those things are options that you can, typically, like, select from, which requires you to know an awful lot of algebra in order to make a, like, a reasoned judgment about what to what to pick from. And furthermore, if you're a seasoned Project Professional, you know that costs are not flat. Okay, it's not like your project starts at its highest level of cost continues thusly all the way through to the end. And then stop spending money. There's typically a curve of code of costs associated like, we'd all love to believe that it's front end loaded, but typically it's back end loaded, where, like, a lot of costs are spent in the, like, last third of the project. So this, I'm finally going to answer your question about how an AI could help. Yeah. So if you're sitting there in front of your computer and you pick, okay, I'd like to do a linear regression analysis against all my costs, it's just going to say, like, current trends will continue forever, and that's your that's your forecast. If you say, I want to do a binomial curve, it'll, it'll sort of taper it off, like a, like a sine wave, almost like off into the, you know, towards the end of the project. But neither one of those may actually be correct based on the data that you have about your actuals and the scope that is left to complete on the project. So you feed those in principle, you could feed all of those data into an AI that would say, like, Okay, so you've done this much. Like, here's what your earned value metrics are telling you you have this much stuff left to do, and here's where your Trend take all that information and synthesize it into a forecast that accounts for all of those variables, as opposed to just one or two of them based on your best judgment as a user. Yeah. Does that make sense? Yeah, it makes sense. So it's but it sounds, and it sounds like a bolt on right, like it's not an existing like we did mention, like Trello, and, you know, there's, there's a bunch of other apps that have it built in. But for capital project, what is, what's some of the software that the individual that we're, you know, you the listener, is using that we you would extrapolate this data from to and then, you know, put it into the AI, and then put it back into like, what are they using? What platforms? So, right now, none. Okay, that's, that's part of what we were talking about earlier, where, like, it hasn't really broken into application space yet. You know, I mentioned earlier a piece of software called forecast that will do this, and it's aI enabled. It seems to be, like, a pretty basic implementation, but it does exist, okay, but in terms of doing, like, the full suite of features that I just described earlier, none. There are none. Okay, so, but they're all in development. Okay? It's like all of the the big names in software, like in project control software that you've probably heard of, like your your hexagons, your innates, you know, contruent, they're all, they're all working on this, yeah. Okay, yeah. So before the end of this year, I bet you're going to see some major announcements from some of those players and others, because this is a key point in in like, the life cycle of this technology, where we're going to see some disruption. Okay, somebody's going to come out with the surprise hit that that out does all of the major incumbent players, you know, I mean, like, that's one of the things that could happen in 2025 is, like, me, know, maybe forecast gets a new version that makes it like, the obvious choice, you know, for doing the kind of thing I just described, and, like, takes your project controls team down from 10 to two. Yeah, yeah. I think it's, I think it's worth, like, doing just full disclosure. So we, like, do condition company is not, is technology agnostic, meaning we're not a reseller of any technology platform out there. And so we, we don't advocate for one of the other. You know, we're not collecting, you. Revenue on, you know, one, because we do do the systems implementations. I think it's just important at this stage, just to throw that disclosure out there for sure. Yeah, like, I've mentioned a lot of pieces of software by name, none of them asked me to, like, they're just based on the research I did for this episode. So, yeah, I mean, we don't get kickbacks, or we don't have license sharing agreements, none of that stuff. So yeah, what is the what are some other use cases that you know, me as a risk professional, that you know for our listeners, might use some of these new AI tools or processes? Well, this is where you actually can start using AI today in a really effective, productive way. And I'll tell you, I've told you this before, Nate, for the benefit of the audience, I think everybody's using AI inside out, okay, like, right now people are saying, like, Well, okay, so they'll come to the AI with an outline, or, like, an idea, and say, write me this thing, yeah. And then, you know, it spews out 750 words of seemingly identical, direct that, sorry, identical, not identical. That's not a word should be, but it's not like if you go to LinkedIn, you can see through the AI, oh, yet, oh, yeah, like anybody who's been using AI to, like, do their little listing image, or write their post, or whatever. Man, you see right through that, like, it's always got the same voice, you know, it's always the same lead in today's market, blah, blah, blah, Yeah, same lead. But yes, getting better at using that. Sorry, they are but like, so that's what I meant when I said, like, I feel like it's inside out because people are coming to AI with an outline, asking it to write the bulk of the text, and then, you know, editing the text that comes out of it that's inside out. To me, I think you go to AI to brainstorm the outline, write the text yourself, and then use AI to proof it. Okay, so if you do that, it has your voice, and you wind up with all of the benefits of AI, like it knows a lot more about what's out there on the internet than you do. That's kind of its whole purpose in life. So it can help you, like, guide your editing process, and it can also help you build your outline. Okay? So if you've forgotten something, you know, while you were outlining in the good old days before AI, you would figure it out while you were writing it, and you have to go back and work it into your outline. You know, you know what I mean. So like, that kind of thing is just not necessary if you have or at least it's going to be greatly reduced if you have a helper bot. So why am I saying this? Because in the world of risk management, one of the most time consuming tasks is actually coming up with risks, like brainstorming risks for a risk register. Okay? Risk Register, fundamental tool of risk management on projects. Okay, so you need to know at least to some degree what your upcoming risks are likely to be, and that's going to be based to some degree on your scope of work. But there are a lot of risks that are held in common across, you know, projects of all shapes and sizes, you know, things like personnel issues and safety lapses, and, material and supply chain disruption, those kinds of things are going to happen to varying degrees across projects of any size or type. So as a risk professional, you can go to any given AI and have it give you a nicely formatted table of every risk that's likely for a project, you know, just give it a handful of basic details, and you can even use some of the the less security enabled ones for this, because you give it generic details about a generic project, and it'll give you back generic risks. Yeah. Okay, so this gives, this leads me to, like my inside out idea what you need to do as a risk professional is take your non secure AI buddy, have it brainstorm a bunch of generic risks for you. Take those away as a risk professional on a project team, hone them and refine them, using your expert judgment, using the data sources available to you, using your cost estimates and your schedule estimates as a guide do your job, but you don't have to do the grunt work. Okay? The grunt work can be done by by by a robot. Why not? Yeah, yeah. And Godzilla is definitely on the list of risks. Yeah, yeah, yeah. If you've ever seen Pacific Rim, it's it's real, it's real. Yeah, absolutely. So you know, if kaiju is not on your your you guys are missing something. You've done something wrong, your force majeure, aka Godzilla. So what is it? So our, let's wrap up, what's your one piece of advice for a risk professional that's sort of looking down, you know, at the AI landscape. How do they get what's, what's your one piece of advice? So I would, of course, is going to be biased towards the way that I like to use AI, but I think, I think I already gave my piece of advice, which is like, start, start thinking about it from, from the other direction. Okay, instead of using AI to to generate the bulk of the thing that you're doing, I'm. You need to give yourself credit for being a professional who is an expert in their field, unlike me, who is apparently not an expert in anything. Just kidding, but it's just a callback directly. Say, if you can remove that, you can say, kill that bit, just cut that out. Safe. But anyway, no, like you treat yourself as an expert and AI as a helper, not the other way around. Okay, so, but once you've kind of gotten over that, like, once you've figured out how what your relationship to your AI tools is going to be, find a nice secure platform, like, for us, copilot works really well because of the Azure Security. Use that to do the kinds of tasks that you're comfortable doing, and start giving it more tasks that you didn't think that you might be able to offload to an AI. Just give it a shot. Like, try. Like, they're getting better all the time. So experiment. Okay, adapt, improvise. Like, do, do what you can to like, make some make some new like, build some new skills using the AI as a guide. Okay, so I guess it's to sum it up succinctly, My piece of advice is, start using AI for the other stuff, like whatever you're using it for right now. Stop using it for that. Start using it for something else. Just see what it can do. Okay, really, really, test the boundaries. Yeah, I like it. Get out there and test all right. Well, thank you. Albert, thank you. Names, a fun one. Yeah, that was a good conversation. So we'll see you all next time. Thanks for listening. Hey everybody. It's Albert here. Thanks for tuning in to the risky planner podcast. We hope today's conversation was informative, and, above all else, inspires you to excellence in what you do. If you liked today's episode, don't forget to rate, subscribe and leave a review. It helps us reach more listeners just like you. I'd also like to thank Thompson Igbo EBO for letting us use his excellent music on our show. If you like what you hear, check him out@igbomusic.com that's E, G, B, O music.com talk at you later you.