In this episode, we explore the impact of Artificial Intelligence (AI) on project management and how it’s transforming the profession. Oliver Yarbrough shares how project managers can leverage AI to enhance their skills and stay competitive in an evolving, AI-driven landscape. Hear about AI’s impact on Agile teams, how SLMs and LLMs are revolutionizing data refinement, how to balance data security while leveraging AI, and how to treat AI as a key stakeholder in the evolving landscape of project management.
Chapters
02:21 … Meet Oliver
04:21 … What is AI?
05:39 … Will AI Replace the PM?
06:41 … Incorporating AI Tools
08:45 … Finding the AI Capabilities
09:44 … Skills and Knowledge Areas
13:02 … AI and Data Analysis Challenges
15:23 … ChatGPT and Data
16:40 … Human in the Loop
18:08 … Protecting Your Data
20:14 … Contractors Using AI
22:20 … Kevin and Kyle
23:18 … Impact on Agile Team Performance
26:22 … Fine Tuning and Refining
29:05 … A Large Language Model (LLM)
30:08 … Current Trends in AI
32:21 … AI Component to PM Tools
33:40 … Streamlining Workflow with AI
39:36 … Future Evolutions of AI
43:17 … Contact Oliver
44:58 … Closing
OLIVER YARBROUGH: AI acts as a stakeholder on our projects, and we should treat it like we treat any other stakeholder. That’s very important. Initially, I used to say treat it like it’s a piece of software tool. But with the new advents of these AI agents and AI assistants and all these other things, you really do need to treat it like a true stakeholder.
WENDY GROUNDS: Welcome to another episode of Manage This. where we dive deep into the latest trends, insights, and strategies in project management. This is the podcast by project managers for project managers. I’m Wendy Grounds, and with me in the studio is Bill Yates. And in the studio today we have an incredible guest who is sure to enlighten, inspire, and, I think for me, educate a lot.
Joining us is Oliver Yarbrough. He’s a PMP, a renowned author, speaker, and trainer with a knack for combining hands-on real-world experience with project management fundamentals. His impressive career includes positions with Fortune 500 companies like Lucent Technologies, Staples, and Sprint, as well as successful business ventures of his own.
Currently, Oliver is an active member of PMI, where he shares his extensive knowledge on adapting to AI, deriving value from data, and recognizing AI trends in project management. He has some LinkedIn courses which we’ve taken a look at. They cover everything from leveraging AI in project management to the importance of human strengths in an AI-driven world.
BILL YATES: As Wendy is saying, we’re going to explore a topic that, I mean, if you connect to the news, if you connect to the Internet in any way, probably one of the top trending topics is AI, or artificial intelligence. We’re going to dive deep into that with Oliver, and we’re going to look at it from a specific lens, and that is from the perspective of the project manager. How does this impact me? How does this impact my job and my future? Oliver is going to share some insights with us. He’ll help us understand how to stay relevant. What are some strengths from AI that we can harness? What are some things that we cannot fear? So, Oliver is here to open our minds.
WENDY GROUNDS: Yup, so get ready as we welcome Oliver. Hi, Oliver. Welcome to Manage This.
OLIVER YARBROUGH: Yes, great. Glad to be here.
Meet Oliver
WENDY GROUNDS: Oliver, why don’t you tell us how you got into artificial intelligence? How you took that path?
OLIVER YARBROUGH: Well, I sort of fell into it backwards. So, I did not start off as an AI person. I’ve always been a Curious George. I’ve always been poking my head in, like, “What’s going on here? What’s going on there?” But, you know, I have a project management background. So, I was doing PMP exam prep. I was training people. And that’s how I got in touch like with you guys. And from doing project management, I’m always looking at what’s on the horizon.
So, I started seeing how artificial intelligence was impacting the financial services industry, how it was impacting healthcare. This was maybe six, seven, eight years ago. I’m like, you know what? How much of a stretch is it if it’s impacting financial services? I mean, with things like Robo Advisors and stuff, how much of a stretch is it for it to start impacting project managers? And I was being sort of front-facing, futuristic. It wasn’t quite there yet.
So, from that exercise, I started writing a series of articles that I placed initially on LinkedIn, back when LinkedIn really pushed their articles. And I got traction off of that. That traction led me to, you know what? I smell smoke. And where there’s smoke, there’s fire. I ended up getting, I think it was PMI actually picked me up to speak at a virtual conference. That went well. Got a lot of LinkedIn followers. And that gave me the idea to turn it into a course. One course led to two courses, led to three courses, led to me now, in the last couple of years, ever since the advent of ChatGPT, I’ve now started to branching out, talking about artificial intelligence, independent of project management.
And what I tend to focus on are the practical aspects of it. As opposed to, like, there’s a lot of people who can do software, and they can build stuff and everything. I’m more in how can we use it in a practical way that I don’t need to be a software expert; I don’t need to be super technical. I just want to get it done. And that’s pretty much where I’ve evolved to.
What is AI?
WENDY GROUNDS: I think it was a few years ago we had a guest where we talked about AI, so I’ve definitely wanted to dig into it a little more. What exactly, for those of us who are novices, what do we mean when we say “AI”?
OLIVER YARBROUGH: Well, AI is basically when a machine replicates human thought. But when I mention AI, because most people are not super technical, I try to keep it very practical. So, what I like to think of it is I combine both RPA, which stands for Robotic Process Automation, which is basically automation. It’s “if this, then that.” So, it’s basically within a framework of sorts; right? And then I also include machine learning, which takes it a step above just simple “if this, then that.” It’s the machine learns as it goes. So, if it learns something, the best way of doing something before, it’ll then incorporate it going forward.
And then true artificial intelligence, is when the machine, the algorithm thinks or acts as if it’s human. And I don’t know that we’ve quite gotten there yet. I think that’s the goal. I mean, some are terming it AGI, you know, Artificial General Intelligence. But just in general I just say it’s when the machine thinks and acts like a human. And we’re sort of on that path, but I don’t know that we’re quite, quite there yet.
Will AI Replace the PM?
BILL YATES: There’s one big elephant in the room, which is…
OLIVER YARBROUGH: Uh-oh.
BILL YATES: …okay, “Project managers, alert, alert. Find a new career. Get a new job. AI’s going to replace you.”
OLIVER YARBROUGH: Run for the hills.
BILL YATES: So, let’s go ahead and deal with that first. Should we all be looking for a career change?
OLIVER YARBROUGH: I would say no.
BILL YATES: Okay.
OLIVER YARBROUGH: I think that all project managers should feel very good right now.
BILL YATES: Okay.
OLIVER YARBROUGH: With that being said, I would be remiss if I didn’t say that we need to be aware of what’s going on and need to prepare ourselves. So, in that sense we do need to start thinking about the effect it’s having, and do things to sort of prepare so that we’re not left behind and definitely don’t become antiquated.
BILL YATES: There’s a white paper that PMI put out earlier this year called “Shaping the Future of Project Management with AI.” And one of the quotes is, “At the very least, fluency in the basics of AI is non-negotiable. It must be in the DNA of project managers.” So, we can’t just bury our head in the sand. We’ve got to kind of go, “All right, this is a new reality. Let’s deal with it.”
OLIVER YARBROUGH: Absolutely. Lean into it; right?
BILL YATES: Yeah, yeah.
Incorporating AI Tools
WENDY GROUNDS: So, project managers, how can they incorporate some tools into their existing project management workflows? What are some ideas?
OLIVER YARBROUGH: That’s a great question, actually. I like to think of it as building your own AI project model. And that’s actually what I’ve called it, and I have a course about that. But the whole point is, from a practical standpoint, is doing it sort of in layers, almost like building a house. You start off with defining your requirements. And what I’m talking about is defining the requirements of what you need to actually accomplish, and so then you can match the various tools and techniques that help you accomplish that.
So, once you have defined your requirements, which is sort of your base, then the next level up is to choose your cloud platform. When I talk about cloud platform, I like to use something that’s more AI friendly, that might have some AI already built into it. So, I’m talking about things like Google Workspace, Microsoft 365. These are two of the more popular ones. Zoho and some of the other platforms have options, as well. But for most of us who work in a business or corporate environment, we’re probably using one of those two I mentioned.
And then the level above that, the third level, would be to integrate your software. Preferably software that has AI built into it already. So, if we’re using Microsoft 365, or we’re using Google Workspace as our base, then we might be able to tie in things like project management software such as MS Project or JIRA or any number of things to accomplish what we want. So, I like to include things like project management software, our dashboards, and business intelligence, as well as any other specialty software which we want to use.
And then the fourth level is when we tie in APIs. So, this is where we’ll have specialized applications. For example, if we want to parse our meeting minutes, if we want to collect data and stuff like that, there are specialized software to accomplish that. So those four levels are the basis of the AI project model. And then, based on our specific situation, we can modify that and use various software to help us accomplish that.
Finding the AI Capabilities
BILL YATES: Yeah, that’s good. You know, as a project manager we have a sense for what we’re good at. You know, here are strengths that I have; here are some weaker areas that I have. Maybe even for my team, here are some strengths the team has, but here are some weaker areas. So, it’s almost like, all right, I’ve got a team. Now I’ve got AI as an extra team member.
OLIVER YARBROUGH: Yes.
BILL YATES: How do I find those weaker areas that we don’t cover, and how do I find the AI capabilities where I can make up for those weaknesses by using AI tools? Is that a good way to look at it, or is that a bad way to look at it?
OLIVER YARBROUGH: It is. I tend to look at, and I’ve actually said this, AI as a stakeholder. So, AI acts as a stakeholder on our projects, and we should treat it like we treat any other stakeholder. That’s very important. Initially, I used to say treat it like it’s a piece of software tool. But with the new advents of these AI agents and AI assistants and all these other things, you really do need to treat it like a true stakeholder.
Skills and Knowledge Areas
WENDY GROUNDS: So, what are some of the skills and knowledge areas that project managers should focus on if they want to stay relevant in AI?
OLIVER YARBROUGH: That’s a very broad question, but I tend to want to lean into more of the people skills. So, when we talk about various knowledge areas, I’m thinking more of stakeholder management. I’m thinking more of communications. A lot of the soft skills. And why do I say that? Because it’s about complementing artificial intelligence than competing against it.
To understand where we should focus a lot of our time and energy, we need to understand what AI is really strong at. AI is strong at collecting data, parsing through it, coming up with the summarizations and the best practices, and also picking up on patterns from large data sets, which are so much larger than any human could possibly parse through. So, if we know that that’s AI’s strengths, then we have to ask ourselves, where are some of its weaknesses? Where can we sort of plug in and add the most value? And that, actually, it might sound kind of funny, but being human.
BILL YATES: Mm-hmm, absolutely.
OLIVER YARBROUGH: Now, what really separates humans from almost anything else? It’s our emotional intelligence, our ability to really, you know, experience the human element. So, as I stated earlier, I like to think of it as, you know, calling ourselves more of a project translator, where we sit in between the leadership, the project team, and artificial intelligence, which we call stakeholders here, agents, you know, assistants, whatever term which we want to actually put to it.
But being a translator, similar to a translator in any other country, we take what one of the stakeholders says, we figure out exactly what they want and need, and then we translate that to the other stakeholders. And by doing that, I think that that’s something that can really help us stand out and really grow going forward. That also ties into both of the knowledge areas of stakeholders as well as communicating with all of our other stakeholders.
BILL YATES: That’s good. One of the themes that comes up frequently in this podcast is the idea of the role of a project manager as the hub of communication. It’s like, okay, here’s the wheel. There are all these spokes that are coming off that are different contributors, or they could be consulting companies or other contractors, even the customer. And that project manager role is right there in the center.
So even though one of those inputs is AI, project manager needs to look at it and say, okay, who needs this data? Are there pieces coming from this AI that are maybe not quite relevant to my team or to my situation? How can I refine that data better or interpret it better for other stakeholders? Yeah, I like that a lot. That’s some kind of an interpreter hub of communication, looking at that and seeing AI as yet another stakeholder providing information. And I love that you say, you know, we have strengths that AI does not have.
OLIVER YARBROUGH: Yes.
BILL YATES: You know, AI’s not sitting here looking across the table at Oliver going, “He’s very calm. He’s very relaxed. He’s smiling.” This news that he’s getting from me on this project update is positive to him, versus if I type the text or just have it interpret the meeting notes, so to speak.
OLIVER YARBROUGH: I like that term “project manager as a hub.” I mean, I think that you may have coined a new term. I love it.
AI and Data Analysis Challenges
BILL YATES: Oliver, one of the things we see from AI, an obvious strength is being able to get useful insights from the project data. You know, as you say, the more accurate data we can feed into these models, the more powerful the data is coming back from AI. What are some common challenges that we see when we’re using AI to do that data analysis? What are some rookie mistakes that you see?
OLIVER YARBROUGH: Well, I don’t know that this is necessarily a mistake of project managers as it’s more or less the inefficiencies in the systems of which we currently use. So, what do I mean? I mean, okay, so again, understanding what artificial intelligence is and how it operates. It operates from parsing large groups of data and then summarizing it and finding out patterns. But in order to do that, it needs the data to be organized and structured in such a way where it can easily go through it. How many of us are dealing in situations where all of our data is centralized, it’s organized, it’s structured, it’s been, you know, parsed?
BILL YATES: Sounds like a dream.
OLIVER YARBROUGH: It’s in spreadsheets. It’s in columns.
BILL YATES: It’s over here. It’s over there.
OLIVER YARBROUGH: No, it’s in somebody’s laptop over here; it’s in some data center over here. It’s all over the place. So that doesn’t really make for the most efficiency when dealing with AI and these large language models, you know, and various other things. So, I would say the biggest challenge is getting organized with our data. And that may sound easy to some, but how many people even have an organized closet at their house, let alone organized data?
BILL YATES: Yeah, yeah, yeah. Just thinking, oh man, something really practical like the AI model may be saying, “Yeah, I’ll give you a good estimate, but can you provide me all your data on your resources from the last three projects?”
OLIVER YARBROUGH: Yes.
BILL YATES: Time sheets, time tracking, that kind of thing. Some companies, they’re like, “Oh, yeah, I got that all day long.” Other companies, “No, you know, we’re not checking into our fast-food restaurant and punching a clock. We don’t do that at our shop.”
OLIVER YARBROUGH: Exactly.
BILL YATES: It’s like, okay, I get it; but, you know, how is it going to estimate if we don’t really have good actual data from those past projects to know how many resources, how long is it going to take? It’s asking for good data, which we may or may not have.
OLIVER YARBROUGH: Exactly. And then how are we going to train our AI assistant, stakeholder? It has to have a pattern.
BILL YATES: Yup.
ChatGPT and Data
OLIVER YARBROUGH: Even if we’re talking about, like, a popular ChatGPT; right? It’s still making a prediction based on the larger data set it has. So, the only difference nowadays, well, one of the only differences nowadays versus years ago is just access to more data.
BILL YATES: Yeah.
OLIVER YARBROUGH: So, a large language model, for example, like a ChatGPT, just has access to so much more data. But if that data is not really structured or if the ChatGPT, if we haven’t trained it on our specific data, which is a whole ‘nother conversation; right? Coming up with, say, a custom GPT or something like that. Basically, to make a long story short is we want to get to a point where we can have a specialized data set that we can then train our own custom chatbot, and now it can give us specific information based on our specific data sets as opposed to just the World Wide Web or something of that nature, which the key point of doing that is we reduce the infamous hallucinations.
BILL YATES: Yeah, right.
OLIVER YARBROUGH: That’s where the ChatGPT or the chatbot just makes stuff up. It sounds good. It sounds intelligent. But guess what? It’s completely made up.
BILL YATES: Yeah, yeah, yeah.
WENDY GROUNDS: Yeah.
BILL YATES: I’ve had some attorney friends tell me, you know, “Oh, yeah, yeah, no, I can’t use AI. You should see the stuff they put out. You fact check, and you’re like, where’s this coming from?”
Human in the Loop
OLIVER YARBROUGH: Oh, yeah. I like the term “human in the loop.” It still requires someone who has some common sense. You know, for example, okay, outside of just artificial intelligence, how many of you have used GPS?
WENDY GROUNDS: Yeah.
OLIVER YARBROUGH: How many times has GPS, as much as it’s been around, and as good as it is, how many times has it sent you down a path or route that just wasn’t right? Or you know, hold on, I drive this route on a regular basis. Like what’s going on here. Things happen. It requires a human being to be able to see that something’s not quite right and get things back on track. I think that that’s a real value that a human being can add.
BILL YATES: That’s true.
OLIVER YARBROUGH: And so, taking the same example, to say a project, you can get a lot of suggestions, a lot of advice. But it still takes that project management common sense to look at that and say, you know what, something’s a little bit off here. Let me tweak this or tweak that and get things back on track so that we don’t end up having a situation where we end up having egg on our face or being embarrassed because we followed something that just did not make sense.
BILL YATES: Yeah, there was a – this is terrible. But there was an “Office” episode where Michael Scott, and I think Jim’s in the car with him, and Michael Scott is driving, and he’s following GPS. And if it says turn right, he turns right. He drives him straight into a lake.
OLIVER YARBROUGH: Wow.
BILL YATES: You know, it’s all for fun.
OLIVER YARBROUGH: Yeah, yeah.
BILL YATES: Sometimes I think, if we put on blinders, we’re going to drive our project straight into that lake.
OLIVER YARBROUGH: Exactly.
BILL YATES: So, yeah, common sense is good.
OLIVER YARBROUGH: Absolutely.
Protecting Your Data
BILL YATES: But here you bring up something that I was looking forward to talking to you about, security concerns. You know, I’m thinking, all right, I’m a company. I’m a large company. We have IP, and we have data about our specific projects, and we have a specific approach that we take. The last thing I want to do is get that data out there into the public space. But I want to feed this stuff into AI. So how do I protect it? I want to make AI smarter, but I also see the other side of that coin, which is I don’t want to put this data out there where it could be hacked or accessed. So where are we today, you know, in this ongoing battle with security and AI concerns?
OLIVER YARBROUGH: You’ve just come up with some of the best questions. That’s been a big concern for a lot of people. I think that’s why the advent of a tool such as Microsoft Copilot has been really good because they’ve really built part of their reputation on having very secure data, and it’s not shared with the larger ChatGPT model. So, you know, again, for people who are already in the Microsoft universe ecosystem, it makes a lot of sense to just add that in.
And the other cool thing about it is, since it’s on the Microsoft platform, and most of us who’ve used Microsoft tend to trust it, and it’s been one of the more trustworthy platforms out there, that’s one way I’ve seen, especially larger companies, enterprise-sized clients deal with the nature of keeping like the data safe and secure and not – the big thing is you don’t want your proprietary information shared with the larger data set, where somebody, or particularly your competition, could just do a prompt and get all your information.
BILL YATES: Right, right.
WENDY GROUNDS: Right, right.
BILL YATES: That’s so good. So, I think this is one of the most practical things that we can say to project managers is just be aware of this.
OLIVER YARBROUGH: Absolutely.
BILL YATES: Ask your IT department, ask your security department, hey, what can we do? Here’s what we’re trying to accomplish. What are best practices? What’s safe? What’s allowed? I want to keep my job tomorrow, so I don’t want to put this data somewhere where I get fired tomorrow.
OLIVER YARBROUGH: Yeah, yeah. Don’t want that.
BILL YATES: Don’t want that.
Contractors Using AI
BILL YATES: Oliver, we were thinking it’d be an interesting thing to bring up to the project managers. What about contractors? Are there any concerns that we should have as a project manager? We have contract firms that may or may not be using AI, as well.
OLIVER YARBROUGH: Well, that’s another great question. My concern with contractors and their use of AI is not necessarily the use of AI. Because if we’re going to treat AI like a stakeholder, then we aren’t going to say, “Well, Bob did X, Y, Z.” It’s just that here’s the end product; right? My concern is less with AI doing work, as long as it sounds good, it looks good. It’s the fact that it may be shared with the larger LLM dataset. I would want to make sure, particularly information that is secretive or more personalized to me and my company, that it’s done through a custom GPT, or it’s done in a safe environment that’s walled off from the larger dataset.
And from a security standpoint, I think companies hiring contractors should state in their contracts that, very similar to NDAs and everything else, that we don’t want our personalized information shared with anyone we have not pre-approved. And that includes AI, bots, agents, anything of that nature. So, putting it specifically in your contracts that we do not want our data shared without our approval, our express written approval in advance, that you can train and/or use our personalized data with, like, a ChatGPT.
BILL YATES: I’m so glad you mentioned NDAs because that’s exactly where my brain was going. It’s like, “Yep, that’s not in there specifically right now, so we can make it specific.”
OLIVER YARBROUGH: It needs to be. The other thing, too, and this is an aside, I guess, but I’m dealing with this now, is putting in stipulations in contracts that you cannot use my name, likeness, or voice, or any, I mean, like a whole sentence without my express written approval in all my contracts, because they will take your voice and license stuff, and next thing you have like a Bill bot or something. Who is that person?
BILL YATES: Yeah, mm-hmm.
Kevin and Kyle
KEVIN RONEY: Across industries, the use of agile project management approaches continues to grow. The Velociteach 2-Day Agile Certified Practitioner Exam Prep class promises top-tier learning experiences and offers real-world applications of the Agile approach.
KYLE CROWE: The instructor-led sessions and online courses offer an in-depth exploration into Agile’s history, methodologies, and strategies—all the key areas required for exam success. Tailored to diverse learning preferences, our course materials cater to auditory, visual, and kinesthetic learners to ensure you will be well-prepared for the PMI-ACP exam.
KEVIN RONEY: The 2-day instructor-led PMI-ACP class is delivered virtually – no travel needed! For group classes, you can choose between face-to-face instruction with an onsite instructor or virtual delivery for private groups.
KYLE CROWE: The Velociteach 2-Day Agile Certified Practitioner Exam Prep class is focused to equip you with the knowledge and techniques you need to pass the PMI-ACP exam.
Impact on Agile Team Performance
WENDY GROUNDS: So, another thing we wanted to ask you, Oliver, was the impact on team performance of using AI for Agile project management. What have you seen?
OLIVER YARBROUGH: Oh, Agile project management. Very similar to other aspects of project management. There’s nothing particularly, you know, special about how AI treats Agile; right? But the thing about it is this. It can really help with coming up with the initial documents, the brainstorming, coming up with, say, the artifacts or sprint planning and things of that nature because, like I said earlier, it has access to, you know, hundreds, if not thousands, of examples of that. So, it can pull out the best practices, say based on what you fed me. That’s what I tend to do.
So let me back up a bit. What I like to do, especially if we’re dealing – and that’s what most of us talk about. Most people when they talk about AI, they’re talking about something like ChatGPT.
BILL YATES: Yeah.
OLIVER YARBROUGH: So, in this example here, you want to really get the most out of your ChatGPT bot. And what I do is I start by feeding it examples. So, one way of accomplishing it is just to do it like a prompt. Just “Please give me xyz.” But a better way, in my opinion, of doing it is to feed it examples of previous, say like a sprint planning.
BILL YATES: Yeah.
OLIVER YARBROUGH: And these are plans of what you’ve done or your team has done, and so you feel very confident in those. And now you’ve sort of trained it very briefly on what you expect and what you think is a good example. And then you ask it, you know, based on these criteria I’m giving you for this project, to then create me a custom, of sorts, sprint plan for this project. Then through a process of refining and reprompting, we get a better example of our sprint plan.
And then ultimately, we can put our own little fine touches on it, you know, personalize it. That’s in general what I would use it for. So, the upfront brainstorming, artifacts and documents, then helping us with things like our meetings and summarizing them, even in collecting of data.
Now, with that being said, you know, humans in the loop are still going to have to help to remove roadblocks, helping with emotional things, folks have issues and problems or whatnot. So, there’s still very much a need for that Agile type of coach or Scrum Master to step in and add value above and beyond what just planning and brainstorming do.
BILL YATES: Oliver, that’s so good. And I agree with you. I see AI as being a great way, especially early in the project, to get some of my initial planning done, whether it’s predictive or adaptive, whether it’s Waterfall or Agile. And to your point, don’t just say, “Hey, can you give me an example of a project charter?” Uh, yeah, you know, here’s generic. Here’s your vanilla ice cream kind of thing.
OLIVER YARBROUGH: Yes.
BILL YATES: But yeah, here are some examples of past project charters that we’ve used that have been successful. Now here’s what’s unique about this project. Can you create one for me? And adding “please,” that was a nice touch, by the way.
Fine Tuning and Refining
OLIVER YARBROUGH: Yeah, I mean, if we’re treating it like a true stakeholder.
BILL YATES: Yeah, like it’s a stakeholder, yeah. I like your idea of using this early on to kind of get my, all right, this is my first draft or first pass. I’m not taking it and running off with it. I’m taking a look at it and going, “Yeah, that’s close, but I could refine it and make it better. Let me follow up, follow up, follow up.” To me, that’s the right way to use AI and a way to use things like ChatGPT is to continue to refine, and improve and improve and improve.
OLIVER YARBROUGH: Yeah. I’m going to actually take what you said a step further. I know you mentioned PMI earlier, and they’re obviously like a great resource.
BILL YATES: Yeah, yeah. Sure.
OLIVER YARBROUGH: I like to use – and this is like a little secret here; right? I like to use a combination of LLMs. So maybe start off with, say like a ChatGPT, feed it and prompt it and all that. But then take what ChatGPT gives you, and then take it over to PMI Infinity.
BILL YATES: Ah, yeah, yeah, yeah, sure.
OLIVER YARBROUGH: Which is – it’s more of a custom type of LLM that’s customized specifically on project management, and more specifically on the large treasure trove of data that’s specifically PMI.
BILL YATES: Yeah.
OLIVER YARBROUGH: I mean, now you can get it really honed. And so now not only are you not starting from scratch, you’ve already honed it through ChatGPT. And now you bring it over to a project management-focused database, and now you get it even better.
BILL YATES: Yeah, yeah, yeah. And you’ve got IPMA, you have Prince, there are so many different – yeah, yeah, yeah, that’s great. AI is not one thing; right? There are so many different variations and so many different data sets that we can use as resources. So, I love that approach. That’s great.
OLIVER YARBROUGH: Yes, and that’s also, too, like we’re taking it from like a general LLM, and then we’re fine-tuning it through a series of specialized LLMs, or actually small LMs. I mean, I don’t want to get too technical. But the point here is that through this series of refining it, we can really get something that’s really, really tight and really, really good. And then, if we add our own experience and personalize it, it makes it that much better.
BILL YATES: Like to me, an obvious example of this is a risk register. Why not use AI to help take a first pass at what that risk register should include from a brainstorming standpoint? Now, again, you have to be careful as a project leader that you don’t let your team off the hook. You know, this is like another team member. This is another stakeholder. They just happen to give us a bunch of ideas that we can start with.
Now let’s see which ones are relevant, see which ones are missing from what we know about the customer or what we know about the situation, the environment we’re going to be working in. Yeah, why not use that to kind of boost your brainstorming that you’re going to do?
OLIVER YARBROUGH: Absolutely. I mean, it’s like a team member with a really big brain.
BILL YATES: Yeah.
OLIVER YARBROUGH: Man, you have a big brain.
A Large Language Model (LLM)
WENDY GROUNDS: What is an LLM?
OLIVER YARBROUGH: Oh, great question.
WENDY GROUNDS: Sorry, I just…
OLIVER YARBROUGH: I try to simplify things, but…
WENDY GROUNDS: … I have no idea what an LLM is.
OLIVER YARBROUGH: That’s a great question, actually, because I’m sure that most people are probably asking the exact same thing. It stands for Large Language Model. So, an example of that is OpenAI’s ChatGPT. ChatGPT is a large language model.
It’s trained on probably millions, if not billions of the datasets. So, it’s sort of general. As opposed to, I also mentioned like another term, a small language model. So, a small language model as opposed to a large language model is trained on a very specific group of tasks, such as like your PMI’s Infinity. It’s trained on just project management data. It’s not trained on HR and cybersecurity and everything else.
So, in that case, we hope by doing that to get more specific information and to reduce the AI hallucinating, -very, very important; right? Because it can just make stuff up. You’re less likely to have it make stuff up if it’s only trained on a very specific dataset, and it can only pull information from that dataset.
Current Trends in AI
WENDY GROUNDS: So, taking it further, what are some current trends in AI that project managers need to be aware of?
OLIVER YARBROUGH: Now that large language models are starting to become mature, we’ve reached the point now where we’re starting to talk about small language models. So, one of the trends are people training their own small language models.
You may have heard of a term called “custom GPTs.” An example of a custom GPT is you can get a subscription with ChatGPT. It costs you about $20 a month. And you can create your own custom GPT. And if you put the right guards in place, you have to basically go in and tell it, don’t share it. You have to press a button inside of there.
But the whole point is you have the ability to sort of wall it off from the larger dataset. And then you can feel relatively confident in uploading your personal data to it. Then you can just get information that’s very specific to what you need it for. So, I’m starting to see an increase in those, custom GPTs.
We’re also, I’ve been reading a lot about seeing a lot more AI being embedded in a lot more devices. So, like laptops, I think you’ve probably heard of Microsoft has been talking about installing it straight from the factory already embedded in there, your phones, your various other devices coming pre-installed with a small language model embedded in the device. We’re also seeing more advent of using voice and AI. So being able to verbally give commands or prompts and getting back responses verbally, as opposed to having to type everything out.
Now we’ve had these types of abilities before, but they’re getting refined to a point where it’s almost becoming almost like human-like. The ability to talk to it is, if you talk to a human, and they’ve been injecting sort of this emotional element into it, where you start to think you’re almost talking to a human, it can get very, very interesting, if I might say so myself. So those are just a handful. I don’t want to just go overboard. But those are some of the big ones that jump out immediately. But there are so many more.
AI Component to PM Tools
BILL YATES: Oliver, you’ve mentioned some of the integration with some of the popular tools also, and I’m seeing that all over the place. You know, Trello, Asana, some Smartsheet, many of the popular tools that our project managers are using now have an AI component to them.
OLIVER YARBROUGH: Yes. You bring up a good point because, when I initially started talking about this several years ago, I was talking about how you had to have a lot more plugins. What we’re seeing happen, as we see in a lot of other industries, are the big players are sitting back watching how these specialized companies are doing things, and they’re starting to build those capabilities into their main platform. So, for example now, Zoom. How many of us have not used Zoom? Well, you used to have to have plugins in order to do, like, summarizations and meeting minutes and stuff.
Guess what? Zoom has an AI companion built into Zoom now that can summarize your meeting without having to have anything installed. Whereas before, for example, you may have had to use another application such as Jasper or some type of content creation software to get more specific answers around, say, marketing or something of that nature. A lot of your various platforms are starting to build that feature natively into their software and into their platforms, and so you don’t need to have as many plugins. That’s another thing I’m seeing, as well.
Streamlining Workflow with AI
WENDY GROUNDS: Let’s talk a little bit more about streamlining workflows. How can a project manager effectively use AI in streamlining that workflow?
OLIVER YARBROUGH: Yes. So, in streamlining workflows, one of the approaches I like to use is to really figure out, you know, what are all of the tasks of which we do. Say you were on a project, just figure out, you know, is it 50, is it a hundred tasks or whatnot. And then, of those tasks, we sort of apply the 80/20 rule. You know, which of these 80 out of a hundred are things that, you know, are absolutely important?
And then of those 80, which ones do I have to specifically do? I can’t outsource to a chatbot, to another human being. It’s specific to me as a project manager. That might be 20, it might be 30 tasks or whatnot. And so, then we wall those off as things that we have to accomplish, or we must do as the human in the loop.
Then the other 70 or 80, those are ones which we can then figure out, you know, which ones of these are best done by another human being. And then of the ones that are left, which ones can I then offload to a chatbot. Why do I like doing it sort of in that order? Because the first thing you want to do anytime like you’re talking about optimizing a workflow is you want to eliminate the tasks that are not necessary or do not add the most value.
So, if I should not be doing something, then why should anyone else do it? Some things just should not be done even by like a chatbot. So, once we’ve eliminated again, we then figure out which of those tasks which remain are best done by, say, me as the project manager and/or Scrum Master. And then after that, well, what’s best done by another human being on our team. And then, if all those are done, then what’s best done by our AI Stakeholder Assistant agent.
That’s just the general mindset I have. We can always apply different levels of software. And I’m very hesitant about saying any specific type of software unless we get granular because the type of software I might use, say, for meetings and streamlining those might be slightly different than something I use to collect the data from, say, other stakeholders. But with that in mind, I do want to give an example. A lot of us, particularly in the Agile world, or even outside of Agile, but specifically in this case in Agile, we like to have these things called “stand-up meetings.”
BILL YATES: Yes.
OLIVER YARBROUGH: Whether in person or virtually, it’s where we all meet up at the same time, typically in the morning or whatnot. And we have about 10 or 15 minutes where we sort of just figure out like what’s been done, what are our challenges and roadblocks, and what do we hope to accomplish the next day, or what’s coming up.
There’s a way for us to streamline that process because, you know, a lot of times all of us don’t need to be in these daily stand-ups. If I’m a software developer, and I’m cranking away, I know what I’ve got to do, I don’t have any issues, I don’t really need to be in that meeting. That doesn’t mean that the rest of the team members should not be made aware of my status. But being aware of my status and having me in a meeting are two different things.
So how we can do this is rather than have a synchronous meeting, we can turn that synchronous meeting into an asynchronous meeting. So, for example, we set up a Google Form, and then we attach that Google Form to a calendar. And we set up a calendar event like we would any other meeting. So at 9:00 o’clock on Friday morning, we’re going to have our next stand-up.
But instead of us meeting either live or via Zoom or in person, we have everyone must submit their information in advance of that meeting, and they can do it in the way that best fits their needs. So, it can be done verbally. It can be typed in. It can be done a number of ways. We can even use recognition software to pick up information off of a form that’s been written down. The whole point is, like anything else, we need to get it somehow captured and, in a text-based format.
So, we get that information that’s been collected from people in various formats, and then it gets collected into our Google Form. And then Google then summarizes it via ChatGPT. And all this can be done, it can be streamlined and all that. But I’m summarizing something that is not terribly difficult, but does take a number of steps.
BILL YATES: Yeah.
OLIVER YARBROUGH: The point here is at the end of the day, once we’ve done this right, when 9:00 o’clock comes, everyone gets a readout, a summary of how everyone else is doing, their challenges and struggles. And you as the Scrum Master, for example, can then look at that and tell, you know what, it looks like everyone’s doing pretty good. The stakeholders on a need-to-know basis get the status updates.
And then the people who are struggling, you can then schedule one-on-one meetings with those people or groups of people, and now you don’t have to involve the entire team on every single thing. The need-to-know basis saves so much time, and it keeps people working. And 15 minutes every day adds up.
BILL YATES: Yeah, it does.
OLIVER YARBROUGH: Adds up really quickly.
BILL YATES: Especially times every team member. That’s good. That’s a good example of automation. And I like some of the high-level approach that you describe with AI, which is, the question is how much human interaction does there need to be? If it’s low complexity, it’s low human involvement, it’s more automation, we can hand it off to AI.
Then you get to the medium. You get to the more complex. Then AI is more of a helper, an assistant, a fact checker, maybe doing some initial research for me. And then I take a look at it and see exactly how does it apply to my project environment. So more and more human intervention, the higher the complexity goes.
OLIVER YARBROUGH: Yes.
BILL YATES: I like that approach.
Future Evolutions of AI
WENDY GROUNDS: Looking forward, how do you see AI evolving in the future, especially for project managers?
OLIVER YARBROUGH: I see AI being more and more embedded in the software we’re already using. I see us similar to what I stated earlier. We’re going to be using AI more as a team member on our team, a true team member of which we’re going to be interacting less and less with text over time and more and more with voice. So very similar to how I’m talking with you now, we’re going to be talking with our AI Copilots of sorts in the future, and we’re going to be able to parse out and assign a task very similar to anyone else.
So, if I have five team members, then the sixth team member is the Copilot. And as we’re giving out tasks, or we’re, you know, doing our stand-ups, we can then assign tasks in a very natural language to our Copilot. And it goes off and does its thing and comes back with the answers and all. That’s one of the biggest things I see on the horizon. Probably I would say we would see this being more common practice, if I had to guess, in the next three to no more than five years. This will be quite common. It will be as common as social media is now.
So, the example I’ve seen and looked at is almost everyone has a social media profile now; right? And that’s pretty common. Twenty years ago, that was not common. Twenty years ago, if you had said everyone is going to have a social media account, multiple accounts, possibly multiple accounts on the same social media site…
BILL YATES: Right.
OLIVER YARBROUGH: …and we’re sitting here managing five, six, 10 accounts in the next 10 or 20 years, most people probably would have laughed 20 years ago. So, in 2024, looking forward, very similar to how we have social media accounts now and various things, I could see a time in the next three to five years where each of us has at least one, if not multiple, AI Copilots. Where I would encourage is the sort of, and where I see the next evolution probably after that, if I’m just being way out there, is having a sort of AI vice president, COO, coach, like your AI Copilot, that’s the Copilot of Copilots, that sort of sits next to you and manages all of your other Copilots as sort of a master Copilot.
Again, that’s going to be probably five-plus years. Is this too far off? I don’t know. I really don’t think so because, as much as people have started using ChatGPT now, the technology is going to streamline itself; whereas right now it’s not as intuitive. It’s not terribly difficult, but it’s going to be as intuitive as setting up a social media profile. You’re going to spin up a Copilot like it’s nothing. And you’re going to have this thing, and it’s going to focus on certain tasks.
So, for example, you might have an AI Copilot that’s focused specifically on marketing, or maybe even a Copilot that just does copywriting. All it does is copywriting. The reason being is it can get very good. It can become a specialist. And it can reduce any errors because all it does all day long is copywriting. Then we have another Copilot that all it does is the budgets. And so now we have a dozen or more Copilots out there. And then they’re all, again, five-plus years from now managed by a master Copilot which works as your right-hand agent. I think this is what we’re seeing. So, is that too far off? Is that too far? Is that like mind-blowing right there?
Contact Oliver
BILL YATES: It’s mind-blowing, yes, yes. Oliver, this is such a good conversation. And the fun thing is, too, it’s just going to continue to evolve. I know you’ve got great resources out there, LinkedIn and other places. Tell us a couple things. How can people find your resources, and how can they get in touch with you to ask more of these questions?
OLIVER YARBROUGH: The easiest way to get in touch with me is through social media. Primarily, I spend a lot of my time on LinkedIn. So, you can always find me on LinkedIn, @oliveryarbrough. And as far as educational resources, I have several courses specifically on artificial intelligence on the LinkedIn Learning Platform, which is where most of my courses live. And if you have a LinkedIn subscription, you can access that. There are other ways of accessing it through your library and maybe even through your alumni association at your college. So, you don’t always have to pay, but if you don’t have access to those, definitely can find them just by getting a subscription and maximizing your use of that platform. And then you can always just drop me a line. And if you have additional questions, I’m always open. I love talking about artificial intelligence.
BILL YATES: I can tell. Oliver, thank you so much again, both for being here in the studio with us – it’s so nice to have, when we’re talking about AI, we have face to face.
OLIVER YARBROUGH: Yes.
BILL YATES: We can read all the elements of communication. That’s just – there’s good irony there. But the contribution you’ve made both to project management and specifically to AI within project management is noteworthy. And we feel honored to have you in the room with us and sharing this knowledge with our listeners. So, thank you so much.
OLIVER YARBROUGH: Thank you. I appreciate it.
Closing
WENDY GROUNDS: That’s it for us here on Manage This. Thank you for joining us today. You can visit us at Velociteach.com, where you can subscribe to this podcast and see a complete transcript of the show.
You’ve also earned your free PDUs by listening to this podcast. To claim them, go to Velociteach.com. Choose Manage This Podcast from the top of the page. There you’ll find a button that says Claim PDUs, and you just click through those steps.
Until next time, stay curious, stay inspired, and keep tuning in to Manage This.
Leave a Reply