Glossary, Dr. Lundqvist, M.D.: Welcome and thank you for joining us for this first in a series of events on AI and the future of project management. By the number of signups for this event, we can tell that many are intrigued about the potential of AI and project management. And frankly, so are we, and we have been looking very much forward to this. But before we get started, let's just do a brief introduction. My name is Rasmus Rytter, I'm heading up our strategy execution department, and I have a very strong passion for both project and benefit management, including designing projects to create value. And with me here today, I have my great colleague and AI expert Nikolaj. Nikolaj P. Hello everybody, my name is Nikolaj and I'm leading a team of AI engineers and AI strategists here at Implement. Together with colleagues and clients, we have been experimenting with AI and its applications within project management. And we have been looking very much forward to sharing some of that work and some of those learnings with you today. Before Rasmus takes over the agenda, I just want to mention that after the webinar, you will be receiving an email with the presentation and a link to the webinar recording. So Rasmus, take it away. Thanks a lot. All right. First on the agenda is AI's potential in project management. And the big question here is, of course, are there any problems in project management we need to solve where AI could help? And without giving away too much of the content in that section, I can tell you that the short answer to that is yes. Then Nikolaj will come back and share a little about where we can use AI in project management and what it takes. As you know, AI is not some sort of magic wand that will fix everything for us. So Nikolaj will help us get a perspective on what AI can do and what we need to do to succeed with getting real value from the use of AI. Then, of course, we'll discuss where to start. There are lots of use cases, but a good place to start is, of course, where AI can really create value. But more on that later. And then last, but really not least, we'll do a demo of our AI teammate. You'll get a sneak peek into the future, or at least our version of it. We will demo our brand new AI teammate and show you a bit of the potential for using AI in project management. I hope you're ready for this. We, of course, also hope that you have a lot of comments and reflections and questions that you might want to share. If you do, then please just share it in the chat. Of course, we hope that some of the questions you might have will be covered in what we have planned. But for all the rest, we'll do a follow up in the email we'll share with you after the webinar. And before we dive into the agenda point, the first agenda point, we'll just do a brief introduction to the topic, sharing a few reflections on what the future holds for all of us who are passionate about projects. Should we be scared to be rendered useless in a not too distant future? Well, we don't think so. But we, of course, should think about the future skills we'll need. Actually, we think that the role of the project manager will be more strategic, more focused on human judgment, on leading stakeholders and change and frankly become more value creating, important and more fun. But the role of the project manager is a big topic. So we will explore that in depth in our next webinar called the human edge. But of course, we need to understand what it takes to create value. We need to understand how AI tools work, how AI can help us and what it takes to create value. And this is what this webinar is all about. Now, the first point and the agenda, AI's potential in project management. And I'd just like to start this section off with one of my favorite quotes. And it is, of course, from Ben Flubia, one of the world's leading researchers on why big projects fail. It says, projects don't go wrong, they start wrong. And his point is well known, simple but uncomfortable. Most project failures are actually baked in from the beginning. Poor scoping, unrealistic benefit expectations, political pressure to say yes before we truly understand the project. So back to the key question for this part of the webinar. Are there any problems in project management where we need to solve where AI could help? And the quote helps us in the right direction. There are lots of problems in how we do project design. So that's a good reason to start there. But with that quote in mind, let's have a look at overall how projects are performing. This is from our survey we did two years ago. And the key element of this survey is the question posted on the slide. And it says, overall, what potential of the benefits do you think your organization is realizing in your transformations? And every time I look at the results, I can't really figure out what I think what's worse. If it's the fact that 18% of the 145 people who responded to the survey states that in their organization, they're actually realizing less than 20% of the benefit potential. That's just wild. But it might be equally scary that only 3% of those 145 people state that in their organization, they actually hit the mark. They are actually delivering more than 80% of the benefits potential. In either case, I think we can agree that we have a problem. And the big next question is, of course, why? And so we're going to have a look at what drives this poor performance. And on the screen now, you'll see the five elements that has a significant statistical impact on our organization's ability to realize benefits. And let's start out with some good news because fortunately, we have some of those. And in terms of technical delivery, 81% of all organizations are actually doing pretty good. There might be some bumps on the road, but in the end, we'll get the technical deliverables done. So that's great. What's really not great is benefits and change management. That's where we really start wrong. And so we just don't do the analysis required to really understand the benefit potential or what it will take to succeed with change. Trustworthy estimates on benefits are also a prerequisite for getting any senior person in your organization to take ownership of benefits. So if your benefits estimates are weak, you cannot expect anyone to want to own them. And lastly, at the portfolio level, if you don't have a good understanding of the value that you can potentially create with the projects that you have in your portfolio, and you don't understand the investment that you need to make and change, then you're also in a bad situation. So all in all, we set ourselves up for failure. We can't make good decisions on project level, and we can't make good decisions at the portfolio level. And making wrong decisions about what projects to prioritize, what projects to kill, and what projects to redesign is expensive. Very expensive. But, Negulaj, that was me with all the depressing facts. Now I think we need a little positive AI energy to get us going. Thank you so much, Kazmas. So now we will talk about what AI can actually do across the project lifecycle and what it takes to get going. So let's jump into that. First of all, I just want to say that AI is going to touch every part of how we work in projects. We have listed here six aspects where we think AI will be extra useful. And just to pick out a few that really excites us. So risk and bottleneck identification, matching patterns across huge sets of data that humans just would not have time to look over. Then there's reporting. So drafting status reports, summarizing meetings, turning structured data into a narrative is something where AI can also be super helpful for us. And finally, then there's planning and estimation. So drawing on these huge repositories we have of historical data, AI can produce faster and more consistent plans and estimates. And the last one is actually exactly what you will be seeing some of in the demo a bit later. And of course, some of this is happening already now and some of it is still emerging a little bit. But the direction here is clear. We believe that there's no part of the project lifecycle that AI won't touch within the next couple of years. So you might say, is it just all hype? Is this something we're coming up with? So we're looking up with. And just to prove that it is not, we we brought a few numbers here today. But these numbers are from studies by OECD and and MIT. MIT. And one thing I think is really fascinating about the top two top numbers here is that we have been kind of used to in automation projects that you have to choose between either higher efficiency or higher quality. But what we see with this technology that we have in our hands right now is that we can actually get both, which is to me completely mind-blowing. And I think the foundation why AI is getting so much attention and investments as it's getting right now. And then perhaps not surprising to anyone who have worked with the technology that when we get into context-heavy tasks or content-heavy tasks, then we are really seeing a lot of potential. So these numbers indicate that AI might become a real assistant for us, not solving all of our work, but in collaboration with us as humans. So let's look a little bit into how we can collaborate with AI. So I really like this mental model that a group of Australian researchers have come up with. So what you see here on the vertical axis is AI agent autonomy. And on the horizontal, you see task complexity and risk level. And if we start at the top left at the autonomous execution, these are low risk, low complexity tasks where AI should just run independently. Think about status updates, dashboard reporting, deadline reminders. These are low risk, well-defined and repetitive. And there's really no reason that our project manager, even today actually, should spend time on these. So here we are drafting documents or making resource allocation proposals or meeting summaries. Where it gets really exciting, I think is on the collaborative execution. So here we're moving into the territory of higher task complexity and higher risk level. But what we increasingly see is that AI is actually able to help us out a lot in this space as well. So here we are talking about sprint planning, effort estimates and risk detection. And AI can really do more and more in this space. And finally, when we get to the bottom right corner here, we move into what they call human led execution. This is high complexity tasks, high stakes, it's team conversations and ethical decisions. And we really believe that AI should never step into this domain. This should be human led at all times. So what does it actually take to build AI agents that are valuable and that we want to collaborate with? So we have identified a couple of criteria here that needs to be in place for AI to move from saving time to creating real impact. And the first one here is context, context, context, context, context, context is probably the word I've been saying the most for the last for the last half year. To build AI agents that we want to work with, we need to make them understand the context within we operate. The next one here is workflow. If an agent is to be really useful, it needs to understand our workflow, the methodologies, the methodologies that we use as project managers and that we use within our company. And the final one here is trust. So we need to both trust the output from the from the AI agent. And we also need to trust the confidentiality of the data that we are that we are loading into it. So if three, three are in place, then we will leave you off to a really good start. But just let's revisit my passion around context for one more moment, because when we have been experimenting with this, we actually stumbled upon three specific context areas that you really need to pay attention to when you're building these agents. The first one is company data. So what does your organization know about how it works, its history and its constraints? The second one here is project specific data. So what do we know about this specific team or the stakeholder map and what we need to do? And finally, there's the project management expertise. So this is the methodology that you are following and the frameworks and the everything that is special to your organization that the AI wouldn't know otherwise. So when the AI tool has has access to all three, the output is genuinely useful as an AI project team made. And without them, it's perhaps just a smaller, smarter Google or worst case, AI slob or just noise that you are that you're creating. All right. So where to start? That's the next big question. And I'll hand it back to you, Rasmus, for that. Thanks a lot, Nibai. And yes, that is true. That is really the next big question. And we have kind of alluded to this a few times by now. So it's probably not going to be an enormous surprise that we propose to start with project design. And the reason why we want to start with project design is to improve early decision making, because that is really where the big money is. And early decision making is flawed because benefit estimates are weak and highly uncertain. Change efforts are rarely estimated actually, if at all. Tech estimates are more mature, but often inconsistent, making it difficult to compare across the portfolio. And deliverables are often not linked to benefits, increasing the risk of scope overload. And therefore, we sometimes, well, therefore, we often see that we prioritize the wrong projects. We set the wrong scope for projects and we miss out on benefits and increased costs as a consequence. And of course, we need to do something about that. And this is where our AI teammate comes in to bridge the knowledge gap that makes early decision making flawed. And we almost can't postpone it for any longer. We very soon need to have a look at our AI teammate, focusing, of course, on the project design functionality. And we're really looking forward to sharing this with you. Our first small step to prove Benflubia wrong by helping projects start right. The example we'll be going through is, of course, just an example. It's a large company, energy company with 10,000 employees. The situation is that we think our finance processes are not working that well. And so the project idea is let's save some time and money by implementing a new finance processes and update systems. Now, I think we're ready to see the AI teammate in action. This is the front page of Implements AI teammate. The use case that we'll focus on today is project design. So I'll just click the project designer tab. To design a project, the first thing you need to do is to add some structured information about the project. And the example that we're using is from a company within Energy and with an estimated of 10,000 roughly employees. The trouble that we're looking at is mostly within the finance organization. So I'll just tick off finance within function. And then I have the opportunity to look at the processes within finance. And we believe that most of the potential is within order to cash and procure to pay. So I'll select those two processes. The benefits that we're looking at will probably mostly be cost reduction and effectiveness. But I believe that there will also be some quality improvement benefits. So I'll tick off both boxes. And finally, we need to have a look at the system landscape and the systems that we will believe that will be affected by this. And it could be a BI system, but we probably also need to do some upgrades within our ERP setup. And before I press load, I just want to stress that this is, of course, the implement version. So in a company setup, you wouldn't need to indicate anything on industry and company size. Obviously, the function would reflect the organization's org chart. Processes would affect company processes. And the IT system dropdown would reflect the systems within that company's IT setup. That said, we're ready to press load. What you see here is then a recap of the processes I just selected before. You can select more or deselect to make sure that you have the right processes that we will be looking at. Then the next step is to look at what tasks or sub processes within order to cash and procure to pay will be affected by this. And it will probably be the approval workflow, maybe the cash application and credit tax, invoicing, payment processing, PEO matching. You also need to provide info about who is affected by this change. And in this case, it's finance. So here we have an overview of the teams within finance. And I'll just tick off some of the teams that I believe will be affected by this. Finally, you need to provide some input on what type of initiative is this. And it's probably a system implementation or at least an ERP upgrade. And there's some process redesign and standardization. And then you're ready to press submit. You get a summary of your selections. And then you're ready to move to step two. In step two, you have the opportunity to provide some additional information. That could be documents, for example, an idea description that contains input on the project. You're also able to provide some text input if needed. And then you're basically ready to let the AI teammate design the project. And just a word on project design, because no matter if we have AI assistance or not, we use the benefit map as a format for project design. And now I can see it's ready. So I'll just scroll down. And here you see the benefit map. And for those of you who are not familiar with that, it's a complete overview of the project where you can see the purpose broken down into end benefits. That again, are linked to improvements in performance. That in terms are linked to the parts of the organization that needs to change in order for us to create the improved performance and end benefits that we want. And then again, for each team, we have what competencies are needed to make the change. And of course, a list of the deliverables that will also enable the change to happen and the realization of the benefits. And if we are just taking a small closer look on parts of the project design, you can see we didn't provide a purpose. So it came up with one. And then it has also provided some estimates on end benefits, which in this case is 5 to 10% lower finance operating costs per year. That in term is broken down into different types of performance improvements that are again further broken down. So you can see that part of this is a 15 to 25% faster invoice processing and cash application per transaction. That again is linked to the different parts of the organization that needs to change. And for example, the accounts receivable team applies standard cash application and credit check routines. And the system has estimated this to be a medium size change. We are using t-shirt sizes to estimate the change. So for the accounts receivable, it's medium. You can see the competencies needed to make the change and then a list of the deliverables. And of course, you also have the opportunity to challenge this design. And what we could for example do is to ask it to estimate benefits in DKK make needed assumptions. Press the button and then it will spend a few seconds instead of a few days to give you an update of what could the benefits be. And here you can see it estimates that we might be looking at 2 to 3.5 million improvements in finance process costs due to process consistency and first time right. You can of course do a lot of other changes, but to make everything easy for the project manager, you can also just export this benefit map when you are done with it to be used in steerco reporting or for portfolio reporting or for portfolio reporting. And of course, you also have the opportunity to save this to continue working on it at a later point in time or to get a total breakdown of benefits, deliverables and costs in the planner tab. But that is another demo. So we hope you find that as cool as we do. So what you saw here is first cut where you're getting some of the benefits of the projects. You're getting an overview of the change needed to realize it and you get a first cut on the deliverables needed to deliver it. And when we say first cut, it is of course to stress that we are not in the autonomous execution space here yet where I can just complete the analysis and make decisions. We still need a lot of human judgment, but it helps us along the way. We also saw how context some of the processes KPIs or chart and so forth is really helping us to increase the quality of the output you are getting here. And then if you noticed the tool here is actually helping you through a workflow that we saw in action that is enforcing consistency and scalability when you are doing planning like this across projects. And finally, if we look at trust, we are able to create these solutions in a way that is secure and that deliver estimates with great consistency. And we just take a look at the value we create, then what the team may does is to create a better foundation for decision making. And that is really great because as I said before, this is where the big money is. Just imagine the cost we could avoid by stopping just one bad project or cutting perhaps 5, 10 or maybe even 20% of project scope in a few projects. In a few projects in your portfolio because you have the data and then transparency to make the right decision. Imagine that. Some will say that there might also be some improved efficiency and that is of course true if you're already doing great analysis. But if you sort of look back at the data I presented before, then in a lot of organizations, we're not really doing that. So what we are getting here is much better data quality to make decisions about projects with no extra effort. Nice. So we think that the organizations that will get the most out of AI in project management are the ones that start already now. And we don't mean just go out and buy some software, but it's actually the real world. The work is in building the conditions for AI solutions to work. So start building the context, building trust and embedding it into existing workflows. And we have started co-developing the AI team made here with a small group of organizations who are already shaping what it will look like in practice. And if that sounds interesting to you as well, whether you want to pilot it or stress test it, then just reach out to have a conversation to either of us directly. We'd really love to hear from you. Yes. And finally, we just wanted to share that this is a series of events and our next event in this series is about the project management role in the future of AI called the human edge. It's on September 10th. But you don't actually have to wait that long to get more inspiration on AI because already on May 27th, we have the execution smartwatch that shares more about how AI is reshaping strategy execution. So there's much to look forward to. But Nikolaj, that was a quick 30 minutes. Thank you to everybody who decided to share the morning with us. We will, of course, share the slides and recordings shortly. And if you have questions, as I said in the beginning, please just share them with us. Or as Nikolaj said, if you want to have a closer look at the teammate, just pop it in the chat or reach out. We are happy to continue the conversation. Thank you.