and the next session. All right. Well, welcome, everybody. Good morning, good afternoon, or good evening. We are so pleased to see so many of you joining from around the world for this webinar. So today we will explore the Gen AI First People function and what that means for us. So we will be looking at how we can inspire you, how we can show you what is happening today with AI, and also possibly give you some insights into what's happening in the near future. So today we have, together with me, we have Alexandra from our office in Malmö, and we have Jesper and myself from our office in Copenhagen. So let's start by introducing ourselves. Alexandra, do you want to start? Yes, thank you. Hi, everyone. Nice to see so many people on the call. So my name is Alexandra First. I have 14 years of experience within the consulting industry focusing on the people and organization space. And during those years, I've had the opportunity to kind of see the development within the people function and seeing also how the technology part is becoming of greater and greater importance for the people function. And now also kind of seeing a lot of organizations starting to kind of testing and experimenting with generative AI. So this is a very interesting topic today, and I'm very happy to be here. And with us, we also have Jesper. You're our tech hero here today. Yeah, the tech guru. That's me. My name is Jesper. I have an engineering degree, background in software and product design. And then I spent about five or six years building a software and data analytics company. And then I moved into Implement as a senior AI consultant just a year or one and a half ago. And I'm actually part of a very growing part of Implement that is focused on tech profiles who implement solutions and build software. We're about 100 people by now. And I'm not a specialist in the human and sort of HR area. But I know a lot about AI. And that's how I will hopefully contribute to our conversation here today. Great. Thank you. And there we have yourself, John. Yes. So my name is John Christensen. And I'm a consultant in the PeopleStrat practice we have. I started out in consulting many years ago. But since then, I moved to the other side of the table. So I've worked in HR in Northern Europe and in the US for many years in various global functions and working a lot with the HR transformation, with divestitures, with transformation technology and rewards and processes around that. And I've always had one foot into the technology space, but looking from the business and looking from the HR perspective. And then I joined Implement three years ago in the PeopleStrat practice and work still a lot with rewards and AI and how to rethink and transform the people function. So, cool. So today, we will have three big topics that we will look at during the webinar. First, we will look at the AI wave today and see what is happening today and what are some of the trends, what are some of the entry barriers, and what is really happening in the AI space. After that, we will secondly go into some examples for the people function. So we will share with you three examples. And for one of those, we also have a demo that we would like to share with you. And then thirdly, Alexandra will be talking about what you can do, how you can take the next steps to start using Gen.ai and how to build some ideas and build a project to get started. So, yes. But before we get into that, then, Jesper, do you have something to say about the latest trends? What's happening in the AI space today? Yeah. I mean, there's a lot of stuff happening. I think we could easily spend this whole webinar talking about just the technological developments, but we'll not. So I'll maybe just highlight two of the primary drivers that we see occurring right now and that I think will shape this next year. So one of them being just general model capability. I mean, the intelligence of these models, when measured, has grown from what was basically a sixth grader onto a full-on university degree student as of today. And we are starting to see the first models. And we'll definitely see here in 25 models that can solve PhD-level challenges. So we are seeing models that do far less hallucinations, like making up facts, that are able to solve way more complex challenges. And the cost of them is going down as well. So this is one of the primary drivers that enable new opportunities. The other part that perhaps fewer people are aware of right now is what's called AI agents. So most people are used to interacting with AI through something like chat GPT. So you send a message, you get something back. It's a fairly simple system. AI agents is essentially when you couple together multiple AIs to solve sequentially a more complex task. And this is what I and a lot of my colleagues help a lot of companies implement. An example of that could be, for example, in the HR space. If you have a lot of internal courses or upskilling, like we do, for example, here in Implement. You could have one agent that looks up and summarizes these opportunities. A second agent that looks at all the profiles across your business. What kind of skills do they have? What kind of projects are they working on? You could then have a third agent that matches these and drafts a tailored message to, let's say, you, John, every two months that says, hey, there's this new course that could be relevant for your development journey as an employee. So this would be an example of coupling multiple AI agents that can then perform a more complex task than we could usually do. So right now, I would say that the bottom line of technology is that there's a lot of opportunities. Those are two of the movers. But the challenge right now is not so much the lack of technology, but more so the lack of organizational adoption. All right. That's super interesting, yes, and very interesting with having more AI agents in the same process. So then why would you think, why are organizations falling behind on the technology development? I mean, again, there's multiple factors for sure. One of them is simply that the technology is moving really fast. So the amount of investment that goes into AI is out of this world. If we compare it to, you know, one of the biggest projects ever, the Apollo program in the US, we're right now seeing a bigger investment in AI than was in the entire Apollo program calculated for today's value. So the investment and involvement of the technology is so fast that it is really hard to keep us up as an organization. And one of the challenges is definitely selecting the right use case, because, you know, where do you start? There's a few sort of things to consider, both the organizational anchoring. Do you go top down or bottom up? And I know we'll touch a bit upon this later. But, you know, are you having sort of a grassroots movement to ideate? Or are you deciding from the top which direction and what to focus on? Secondly, when you want to select a use case, you have, you know, a balance of really to understand both the business aspects of things, but also the technology technology to the extent that it's not a magic wand. Like, there is a lot of misconception around AI, and understandably so, like it's evolving very quickly. But if you don't have a thorough understanding of the capability, you'll likely end up choosing a business case that's strong, but a technological application which is just not suitable. Because there's a lot that AI cannot do today. There's a lot of things it can also do, but you need to be very aware of that distinction. And then I think lastly, even if you do, you know, end up selecting the right case, building a good solution for whatever process you're optimizing, then a challenge today is that I see a lot of companies investing a lot on the tech side, and a bit less so on the adoption and training side. And this is a bit of a, you know, behavioral science kind of topic. We need people to change behaviors to recap the benefits. And we shouldn't underestimate how much you need to invest in training and upskilling to get people to change habits. So technology is really only half the coin here. And organizations need to understand that. That's very interesting. Thank you, Jesper. And coming from the tech side also. Yeah. I love sitting and building stuff, but I can see we also need the second half. So then do we also see the, for example, the EU AI Act, do we see that as an entry barrier? Yes and no, depending on what you want to build, right? But I think the most important step is that we actually have a legislation in place now. So one of the barriers has been also the clarity around the legislation and what you can do and what you can't do. But now since we have the EU AI Act, and of course, that's also developing, then we have more clarity. And we can also see, I mean, organizations, our clients kind of maturing in how do they navigate within the regulation that is in place at the moment. So I think that it's that help rather than a barrier. And we can also see that in the examples we will present here today, right, John, how we have also kind of navigated the regulation in terms of what we have built. Yeah. Okay. All right. Well, thank you both for those insights. So yes, so let's move into that. And let's look at what some examples from the people space. And based on what Jesper just said, and also Alexandra, what we have, what we're showing today is really like one agent solutions. So we're not, we don't have any solutions with multiple AI agents. And also from an EU Act perspective, there's four levels of risk associated with different AI solutions. So what we are looking at today is like minimal risk solutions. So we are basically automating or doing some work tasks, but we are not doing anything that would rank or evaluate or assess people and could potentially potentially have impact on employment opportunities, promotions, et cetera. So we're looking at minimal risk ideas and tasks today. So the three examples that we have, the first one we will talk about is employee surveys, how you can do that differently. The second one is really a skills finder or finding the right people. And the third one is a people help desk. And the third example is also something that we will show you in a demo afterwards. So we have a a little bit more detail that we can see on that. So the first one of these, the employee survey. So if you think about it, if you have an employee survey today, then you probably have a scale of one to 10, one to five, and then you have two or three or four questions that you ask people. And then you ask them every month and you get a result like 4.5. And then you discuss, oh my God, we got 4.5 and that's really great. But what is happening is that you're discussing a number. You're not really having a dialogue with your teams or your organization. And I think that's a great example. Can you also elaborate a little bit more on how is that different from the surveys that we use today? Because I think most organizations are used to using surveys nowadays. Yeah. So really what Gen AI, what that opens up from an employee survey perspective is that now you can ask open questions. You can ask, how are you doing? Or what should we do? Or what should we focus on? Or what kind of problems do you have? So you can ask really open questions and then people can give their feedback a little bit or a lot. And then you can use Gen AI to really summarize all of the feedback. It doesn't matter if you have thousands of employees. It will summarize into the key topics. You can also ask it to do like an action plan, communication plan, or to say what should we prioritize in this situation. So you have a lot of opportunities where you have, first of all, you're opening up the dialogue. You're not controlling the dialogue. You're opening it up. And then secondly, you actually have the opportunity to create more of a conversation with your organization rather than having more of a, you know, a little bit of a closed one way way of communicating. So that's the example of the employee survey. Yeah, that sounds really cool. I think that's something that a lot of both people functions, but also actually managers would like to have in their organizations today. And Jesper, you have the solution to one of our key problems, right? We're always looking for the right person with the right skills. Yeah, I mean, so this is probably not uncommon to a lot of organizations. Any organizations that work project-based and have a lot of sort of focus on combining the right skills for the right projects probably has the same struggle that we used to have, which is finding the right people. We are 1,500, 600 consultants around here. And obviously finding the right match for a project is not always easy. Typically, you know, we have an internal database with CVs and what projects we've been on. But historically, the way you have to search these things, you have to hit the right keywords. If people haven't updated their profile, you might not find them at all. And what we did is that we built what we call the AI People Finder, which essentially coupled a new type of AI, enables search around semantic search. So it's searching not just for keywords, but for the meaning of my or your search. And secondly, we coupled it with more data sources. So we coupled it with different project databases that we use to register our projects in. And because of Gen.ai, we were able to extract information from that and format it in a way where it's useful. And this is something that hasn't been possible before. So our entire search function within implement is now AI. AI powered in that sense. And I think any organization that, you know, work on a project basis and need to find and move talent around has these challenges. And this is another great use case of using AI to more intelligently understand text and understand semantics. That's really cool. So you don't need to have those exact words. Yeah, it's all words. And we discussed this earlier, but this is kind of relevant for all organizations that kind of have high skill sets, right? So it could be, you know, project based within construction, marketing, professional services. Engineering company, software. Yeah, exactly. Cool. And John, you want to tell us a bit more about the kind of internal, more HR operational partner we have been working with? Yes, I'll be happy to do that. So we have actually also built an AI people partner. So here the problem we were looking at was, how do you find information about, what is the onboarding process? How about promotions? How does that happen? Or absence, vacation, or how do I do this? So basically, we sat down and we saw the problem was that we have this process that is fairly administrative. It's a little bit cumbersome. And if you implement a traditional help desk, then you have to open a ticket, it has to be recognized. And then somebody has to be assigned to somebody who will look up, you know, so you have a fairly long process. So we were thinking, you know, how could we really cut a solution to the chase and say, minimize the distance between the question and the answer to the question. And Gen AI, of course, can do that. So we trained the AI with our own policies. We also actually trained it with guidance for people managers to say, you know, what would be some good advice to deal with a certain situation, right? So it's actually a combination of just being a standard help desk, as well as also being really a partner, potential partner for your managers and others in your organization. So it has almost a little bit of coaching skills, right? Yeah, that's probably taking a little bit too far. But yes, you can. You get some tips and tricks. Yeah, so you can. It basically, so really what is important is, you know, what should be the focus of your AI solution, right? What is it? What is the problem you wanted to resolve? And then you feed it, you need to have good data quality. And then you feed it and you train it. And then, you know, and then you can use it for that purpose. Yes. Cool. And yeah, you want to say something about how we develop these tools? Yes. I mean, those are all examples of what we internally call dev labs, development labs. So it's a format or framework that we have used both internally, but also externally with clients, being like an agile four to six week sprint method, where we combine both sort of people with deep business insights into a certain topic, and someone with a deep technical knowledge. And then we really iterate on the concepts. And we do a lot of iterations of prototyping, if you will. Because the truth is that a lot of these functions are new, like the use cases are new for an organization. So you need to have an agile way of approaching these new concepts so that you can rapidly test them out. And by four or six weeks, what we have is a proven proof of concept for a given, you know, use case, which we can then choose to mature. But those are examples of frameworks or places where we have applied this framework to really build quite cool solutions, I would say. So what we're going to see now is something we build in six weeks? Yeah, actually less. I think it was four weeks or something. But yeah. Nice. So I guess you're all curious now. So let's have a look. So here comes Marie Hansson from the People's Practice. And she will be presenting the AI People Partner. Welcome to the AI People Partner demo. During the demo, we will guide you through solution developed for implement consulting group. Here we would like to showcase some basic functionalities and real life prompts where employees are asking the People Partner for support on various people related questions. When you open the AI People Partner, you will be guided to the front page. This page highlights the key messages we want to convey to the users of the tool. It highlights that you can easily navigate through people policies and processes, get concise and friendly answers referring to relevant documents, confidentially mentioned sensitive data such as names, and your data is not used to train the model. AI People Partner is generative AI. This means that everything returned by the system is a creative output and not fix. So always read through the output before using it and take everything with a grain of salt. By clicking accept, you will enter the main page of the AI People Partner. Here you will find different features in the menu to the left, as well as the possibility to ask questions about policies and processes. Looking at the first feature, we look at the feedback button. This can be used to bring a feedback button. This will provide any feedback on the tool for any users. For example, I asked how much vacation I accrue and the answer was incorrect. So can you please help update the information to correct the answer? When submitting this feedback, it will then be sent to the People Team for further action. The next feature is statistics. The statistics feature shows usage data such as how many questions have been asked and which topics are the most common searches. Here you can see that within this month, holidays are the most searched topic. To re-enter the homepage, you simply click back again. The last feature in the menu to the left is choosing the country you are in. This is for instance for global companies where the processes and policies might differ between offices and countries. The tool is built so that the documents in the back end of the tool are connected to the specific country in this field. So now we will move into showcasing some real-life prompts, starting as an employee hired in Denmark. The first question we would like to ask is how much vacation do I accrue as a Level 1 consultant? Here it gives a short summary with the answer that I accrue 2.08 vacation days each month. Further, it also attaches a source where I accrue as a Level 1 consultant? As a follow-up question, I would like to know which days are actually paid vacation days. Here it again gives me a short summary with the answer as well as a source. Then I would like to know which days are actually paid vacation days. Then I would like to know how much paternity leave do I have as a dad. Here it shows that a dad is entitled to two consecutive weeks of leave with pay immediately after the birth and additionally he has the right to 22 weeks of leave with pay. Now, let's test out some of the global prompts. So we are choosing all countries in the country field. The first question I would like to ask is what is the global onboarding process? Further, I would like to know what are my responsibilities then as a buddy in this onboarding process? Here it again attaches a source that I can click on to read further details. These were our examples based on an implement context. The AI People Partner can of course be trained to support your most frequently asked questions within your people organization. This was the demo for today. Thank you for watching. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. moon rocket just yet. Not yet, maybe soon. And we have a lot of experts and expertise within this area as well so if you're out there and you're feeling a little bit lost in this space I mean we are also here to help. But let's look at the, yes sorry maybe we just mentioned here but there's a big spread I think as we mentioned but the lack of competence is actually one area. So lack of competence is number one right and then legal aspects number two and then the other areas. But fairly well spread out. There are challenges. Yes there are challenges. So this is great and thank you so much for participating and leaning in with good answers to these questions. So now Alexandra maybe could you share with us you know and now looking at lack of competence is actually number one here right. Yeah. Could you share with us what can people do to get started right. What are some of the steps that they could do could take to move on with Gen AI. Yeah. And as we have kind of discussed here today already to build the right AI capabilities and this goes for I mean the people function but also kind of any organization today as you mentioned Jesper you need to kind of look at it both from a bottom up perspective but also from the top down perspective. So you need a bottom up to kind of build the understanding finding the right use cases the right right areas where you can leverage AI and just testing and trying. But you also need a kind of top down strategy to kind of provide that clarity around where we're going and how are we going to go from that experimental phase to actually integrate and scale AI. So if we start a little bit with the kind of bottom up perspective and this is as you mentioned as well the method we've been using to actually develop these tools that we have here today. So there is a foundational step that I think goes for everyone. It's just to kind of emerge yourself within what is generative AI. What does it mean for HR or the people function. I mean listening to webinars like this one or just reading upon the legislation and so on just so that you feel more kind of comfortable in what it is and also encourage others to do the same. But then if you go to the actual organization and you look at the people function. So the first step is to look at what are the typical kind of use cases the areas where we could use generative AI in our organization. And a good way could be to look at the people processes or you know the typical tasks that you are doing in the people function and see where are the biggest pain points today. What maybe takes up most of the time or where do you have the biggest issues and where do you think is the biggest potential. And then you need to prioritize those. So the prioritization is about where is the biggest impact. And looking at this from a business perspective then we have you know it could be efficiency but it could also be things like quality like you mentioned in the survey that you want better quality. And then we have the technological part of it as well that you mentioned Jesper. Yeah I mean because one challenge we sometimes see is that you might have identified what is actually the biggest business pain and then you go forward with that because the potential is huge. But the problem is back to the magic wand that if you don't really understand AI you think that it can do everything. And there's a lot of you know really great progression happening but there's still a lot of things that AI cannot do. There's a lot of limitations where language models like ChatGBT are just not suitable for the time. So here it's not suitable for the task. So here it's really important to have someone in the organization or hired out from the outside who has a deep understanding both of the current limitations but of course also where things are moving so that you can choose the right sort of case. And often that case is not number one on the business sort of prioritization. Maybe it's number two or three or four. So you have to make that trade-off. So you actually select something that both has a meaningful impact from a business perspective whatever that focus is. But also something that is technically feasible and reliable production ready to do now so it doesn't die as a you know a test that was cool but didn't really work. So I think that's the trade-off when you do the prioritization is these two axes in a sense. So it's really finding that kind of perfect match you could say but also bringing in the right capabilities understanding the business understanding the people function but also understanding the tech part. So now you have your use case. So now you have your use case. So now you have your use case. You have prioritized your use case. And now it's the kind of business case around it. So what are the kind of investment needed to get going with your test version. And now when you have that you have actually a proof of concept. So you have something that you can evaluate and see how is this and this is something we can scale further. And the best thing about this process is that we mentioned before it doesn't need to take that long. Four to six weeks we said. And you actually have a lot of business. And you have a lot of business. And you actually have gained the experience of working hands-on with generative AI. And you have an actual solution that is possible to scale. So you have kind of built a little bit of those capabilities needed. But that was the bottom-up perspective. So then you also need what we mentioned the top-down. So you need to look at the strategy. So what kind of people function do you want to build? What is the roadmap ahead? Where do you see yourself leveraging? What is the roadmap ahead of the top-down? And then you need to look at the top-down? And then you need to find that balance with the bottom-up approach. And the best way is to have that strategy in place, the clarity, but also to give enough flexibility in that strategy to allow for that experimentation to happen. But also to be kind of, it's okay to test. And it's okay to actually fail sometimes to actually find that right solution in the end. And you probably need to do it. I mean, because that's also, I think, where this differs a bit maybe from other sort of change strategies is that the technological field of AI is moving so quickly right now that it's a bit harder to do like a full sort of top-down roadmap for the next three years. I mean, it makes sense still to do some of that, but you really need to have it in a very agile framework because the technology is changing faster than any other technology. So it becomes very difficult if you have a too rigid structure in your organization. You will miss some of these opportunities. You need to have this agile sort of structure in place when you do these kind of things, even from the top-down approach. So you can say, have a direction, but be ready to change it and be ready to experiment. From your experiment and take those learnings into consideration. And also give that freedom to the organization. And don't be afraid to kind of get your hands dirty and start experimenting. Cool. Good. So one of my takeaways from what you just said, Alexander, was also that even if you do the bottom -up and you're looking at your processes, so maybe you have your people processes mapped out with 15 steps, but then try to take those, but then take a step back. And say, where do we have the pain points? You know, what part of it? Or what is really the problem or the task that we're trying to resolve? Right? And then you will find that your typical help desk, for example, it has a good number of process steps. You know, so it might have 10 steps, it might have 12 or 15 or something. It probably has, you know, cannot have less than 10. But think about it. If you can then say, what is the problem or the task you want to resolve? And if you can use AI to that, then you're down to having two steps. And that in itself is just a tremendous gain. So, all right. But thank you very much. I'm just looking at the clock. We are, you know, we are fairly advanced on time. We can see that we got a couple of questions, but they are also fairly detailed. So we will get back to you on the questions, but they are a little bit long and detailed. So we cannot get to those today. But thank you so much for joining. We hope this inspired you. We hope we gave you a flavor of what is happening in the future. Also some concrete examples. And then thirdly, the steps that you can take so that you can move ahead with Gen.ai. Yeah. Jesper? I guess maybe just to wrap it up, I guess the primary takeaway from sort of a tech perspective is really that a lot will happen on the model side. So whatever limitations that you're experiencing today in terms of chat GPT or the like not being able to do something well, that will likely, you know, go away to some extent. So there's really the understanding of how the technology is evolving is important. What parts that will improve because some parts will improve a lot. And understanding, I think most and foremost today that technology is not limitation for 80% of the good use cases out there. So it's really more about sort of focusing on the organization as of today to get the benefits. That's a good point. Yeah. I think it may be. Yeah. Yeah. No, definitely. I mean, super interesting. And as you've been mentioning, the technologies here, but the barriers are really kind of in the adoption and also kind of that behavioral change that we're looking for in the organization to get the people with you to be able to move. And we also have a lot of experience in this field. So don't be strangers. I mean, reach out if you want more inspiration or just someone to discuss with. We are really happy that you listen in today. We wish you a wonderful day and enjoy the rest of your week. Thank you. Thank you. Thank you.