Rasmus, Rasmus. Good morning and a very warm welcome to today's webinar about generative AI in B2B sales. I'm Rasmus. And I'm Tobias. And we are really thrilled to see you. We can see the numbers on the screen here that we have a bit more than 100 people with us and the number is still going up. We thought that it would be great just to get a feeling of who is actually with us today. So if you'd please just check in in the chat and write the company you represent and where you're located. Let's have a look. Yep. It's always nice to see usually a lot of good people checking in from all over the globe sometimes. All right. It's coming in now. We have Denmark. Tobias and I, we are also located in Denmark. We are in Copenhagen right now. But then it's going crazy now in the chat and we have at least Switzerland, Sweden, Holland. What else have we got Tobias? Norway. So Norway, yes. The UK. So mainly Europe is represented. So no very early birds from the US or the like. Or very late. Or very late. Yeah. Exactly. All right. So a warm welcome to all of you and thank you so much for checking in. I think I'll just start with a short little story about the whole intention behind this webinar here. So back in November, I came with full enthusiasm and excitement up to Tobias. I met him in our big canteen here in Copenhagen and I said, Tobias, we're going to host a webinar about generative AI in sales. And Tobias responded immediately, yes, let's do it. And he said, however, I'm a little bit busy. So within the next couple of weeks, I'm actually not really ready to do this. So he asked me, when is the time for the webinar? And I said, no worries, no worries. It's in three months. It's in February. He looked at me. Tobias looked at me. He said, how on earth are we going to do something in February about generative AI? That's like in 10 years from now, things are moving so fast. So how can we do anything? That is true, I guess. And that also means that we have prepared and not for three months. We waited a little bit to see where is actually the state of generative AI in February 2024. So the whole purpose today is not predicting the future at all within the space. It's probably not possible for now. So what we'll do instead is that we'll try to be pragmatic, try to show cases, but also try to give you some practical inspiration that you might find useful in your everyday. So that's the intention and we hope that's okay with your expectations. But first, we'll just do a short intro of who we are. So I'm Rasmus. I started and implement almost 10 years ago. I'm part of the practice or the function area that works with commercial transformations, where we do anything within the space of sales and commercial excellence, commercial capabilities, pricing and monetization, and so forth. Specifically, I'm super passionate about business to business sales, the processes, setting up the infrastructure, working with larger transformations, and also in the field of working with value selling. With regards to AI, I guess I'm a little bit of a newbie, but I guess a lot of us are. I picked this up last year, and I've been trying to sort of set the agenda within the commercial space, and especially within sales, and see how we can benefit from applying generative AI in sales. So I've done a few of the first pilots and projects with clients, and I'm also the must-win battle owner of this concept in my practice. And then I'll hand it over to you. Yeah. Just a few words from Tobias. Thank you so much. Yeah, my name is Tobias. I've been in implement for just over three years, and have worked maybe for the past five years or so with AI. I'm passionate about data enablement. So how do we secure data to be used in AI tasks, but also other tasks in business intelligence and so on? Then I'm very passionate about solving the hard problems, but also the right problems using AI. And I think thirdly, I would mention that I'm really passionate about training actually. So I've done a lot of training both for some of our clients, but also internally in implement, and have done, I don't know, too many sessions to count almost. So that's me. All right. So now I think for the next step, Rasmus, we would love to just get the temperature of this room, just to get a sense of where you all are. Let's do it. So I think we'll try to put up a poll, and we're just really curious here, sort of where you are in the hype cycle, you could say. So sort of ranging from, I cannot see the point of applying generative AI in sales, to I think it's an obvious and natural application of enhancing sales. Where do you sort of fit into that spectrum? It'll help us know our audience a little bit, then we can dive into certain aspects. Yeah. So I think that's nice. And we're getting the results here. Yeah. Live on screen. That's so cool. So the first votes are in. The initial picture is pretty clear. Yes. We're on the end of this sounds reasonable. We'll probably use it. Maybe we are already. We'll come back to that. I think at least we don't have any in the very first category of cannot see the point of applying generative AI. So I think that's a good starting point. But maybe there are some when we get some more. Maybe they didn't join this webinar. That would also make sense. Yes. Yeah. Okay. What do you say? We are seeing some of the votes. Maybe some of them will come a bit later, then we can touch upon it again, if relevant. Yeah. We don't see a lot of votes. Ah, here we go. We've got more than 100 votes. That didn't really change the picture. We might have one person here that is maybe in for a real piece of inspiration. And hopefully we can convert someone to the other side. Let's see. Let's see. I guess it's also okay if not. If not. We'll do our best. We'll do our best. Yeah. Exactly. All right. So thank you for checking in. Let's have a look at what we'll cover today for the next, yeah, just about 40 minutes. We will run a mix of just a little bit of theory, a few perspectives, and then we'll try to look more into the application of it. As I said, we will try to get perspectives from a, let's call it a so-called first mover. We have been interviewing a CSO. So someone who is relatable to most of you, I assume, and that also has some real experience from reality. You'll even show a demo. I will. With full risk of what that implies. Yes. Said the guy who's not in tech. You're really comfortable, so that's fine. Yeah. And then along the way, you know, please use the chat. You can post questions, observations, reflections, learnings, whatever that comes to your mind. We will not have time to respond during the webinar, but we do have some colleagues from Implement who is also on this webinar, and they will do their best to respond if they can. Or else we'll take them, note them, and then we'll come back to you at any time. So I think that sets sort of the scene for the plan here, and we have four simple topics to cover. And the first one is around some of the potentials. But before we look into the real potentials, let's also get a little bit of semantic in place. So we talk about the same stuff here. Yes. Absolutely. So I think when we're doing a webinar on generative AI, it makes sense to just cover the basics, all of it. And I think what we usually like to present is a picture like this. So if we sort of go to the broadest of senses, and we talk about artificial intelligence, that essentially covers techniques and methods for machines to replicate human behavior. This is actually very old. This is an academic discipline. It goes back to the 50s. And even Alan Turing discussed some of this. Within that artificial intelligence outer layer, we have two methods, you could say. Machine learning and what is called deep learning. So becoming a bit more concrete, machine learning is the simpler of methods. It's useful in, you know, Rasmus, on your Gmail. Sometimes you'll get spam email and you'll get good emails hitting your inbox. That is machine learning. It's seeing a lot of data. It's learning those patterns and then saying spam or not spam. That would be machine learning. Deep learning is very similar. And you could also do the spam case with deep learning. However, it's bigger models. They are sort of usually wired seemingly close to how the human brain is wired. And these can be used to then sort of classify photos of products. Or they can count the number of cars on a road or whatever we might want to do. Again, tasks that we as humans could do. We could count the cars. But here we're using computers to emulate that. These are more complex and more data are needed for those types of models. And within this box of deep learning, generative AI also fits in. Because it's a subcategory of AI. But it's based on huge deep learning models. These models are very good. At the moment they are sort of, you know, everyone has a bachelor's degree in every topic almost. So that's sort of how we can picture them. But I think I should say some textbook definition kind of thing about what is it then actually. So generative AI essentially consists of applying algorithms to make new content. So this content excels by being similar to what a human might produce. Just like we could write some text, so can generative AI models. They can also do audio or code or images or even videos is now moving ahead. But again, it's purely technology driven. And with even better models, we're seeing even better performance. And that makes it more usable from a business point of view as well. Yep. All right. I think the real reason that maybe we are also here today and the reason that generative AI as a topic is very hot. Is that there are productivity and quality rewards to reap. So we are monitoring a lot of studies in the field. And so far what we see is that there are approximately 32% increases in productivity and 18% increases in quality. So these are studies based on white collar work. So, but generally speaking, this is what we can sort of reap into our work tasks. And we see, you know, it's lifting the bottom performance. It's increasing employee satisfaction and so on. So these facts are essentially sort of the backbone as to why we are here today. So that's the AI angle at least and the short AI introduction. Yeah. And when we look at it from a sales perspective, there are also being done a lot of research right now. Whether it's Bain, McKinsey, Gardner or in this case, as you have on the screen now from BCG. I mean, they all show the same. There are huge potentials, especially also in the function of sales. So, so again, these are just expected impact potentials that we look at now. You know, we can generate much more leads. We have higher capacity to handle the leads. We will have possibilities to do even better cross and upsellings for bigger deals. And also we are potentially seeing a lift up in the in the win rates or the on the lead to to sales conversion rates. Sounds pretty good. I mean, these are numbers. These are the estimates. And I think we need to sort of break it down and put it into a bit more of a of a practical point of view and also start to see some of the evidence behind this from the real world. And I think just initially, I'll just give you a simple framework to look at it, which is basically this formula that says, you know, if we can improve on productivity, multiply that with also better quality. We can somehow improve our sales performance. And that can then be broken down into two different aspects. So basically, are we able to do things faster? Can we generate content faster, emails? Can we do analysis faster to better understand our our target segments quality wise? So can we have people preparing even better, following up even better, either using the Gen AI or through automation methods as well? Coaching. We all know that sales coaching is a key driver for success that has limits to the to the people of the leaders and their time. So, you know, can we do something with that and thereby increase the quality? Ideally, in some way, this would sorry. Yeah, this would lead to to some sort of impact. And that obviously depends on the use case, on how you would approach it, where you would apply it. And we'll come back to that. But ideally, this should have a potential uplift in we will we will come back to what exactly is it that that we they are doing within generative AI and give you a little bit of a bit more of meat on bone in that regard. But but I think to kick it off also and move forward, I think it's also about, you know, making the decision. So maybe, you know, let's let's see, you know, what is it that, you know, what can we how can we be guided towards these decisions that that Anders says, as an example, was one of the early ones to do? Yeah, yeah, absolutely. And and I think to to to sort of help guide those decisions like the one Anders and his teammate, I think you help with having insights into three aspects or three challenges. So so one is from from a business side in this context of sales. Where do we actually play? What are the use cases we want to pursue? Because we know there's a ton of use cases here, you could apply it basically everywhere. Yeah, but figuring out where do we play? Yes, then there is a tech side. So what does this actually take in terms of technology? What does that playing field look like? We won't go too too deep on the tech today, we'll show some tech, but we won't go too deep. So so so don't worry. And then I think finally, there's a there's a change and an adoption challenge to definitely. So so what does what does it actually take to sort of unlock the full potential of these types of solutions? Yeah, so so hopefully we'll try to give some insights on each of these three challenges. This can inspire you and then help you guide forward to to potentially making that decision of of doing more on generative AI. Yes, so so I think what we'll do is we'll try to get into it a little bit with the the options that lie within B2B sales from a generative AI point of view. And then what we try to do, Rasmussen and I is is to sort of draw the playing field, you could say. Yeah. So where do we actually play? And it's it's huge. There's so much to do with generative AI. So what we try to do here is to to draw up a sort of radar with two axes. So one is sort of what is internal sales operation. This is on the vertical and what is sort of customer facing customer engagement. And then also what do we use to sort of improve our existing processes? That's on the left what we call incremental generative AI. And then on the right what are the new products, the new services, the new capabilities that that we are yet to enable or maybe we don't even know what looks like yet. Yeah. So I think starting from that, starting from the sort of incremental customer engagement point of view in the in the top top left. These are potentially use cases and ideas that can boost our sort of customer facing salespeople. So where could be a lot of ideas out there, but but one that it's in my mind could be could we send a one page brief for a meeting preparation to a salesperson? It will take in on his iPad in the car, his or her iPad in the car, and then he or she is prepared for the meeting. Maybe something about the pains and the value drivers. Better prepared. Better prepared. Let's call it that. Yeah. Also, then if we sort of segway to the to the bottom left in sort of boosting our internal operations, that might look like helping make better, more engaging product descriptions. Could be something about the technical explanation that might be too technical, maybe we want it to be more customer relevant. Yeah. That could be some of the internal stuff that we're boosting. If we move to the to the bottom right and assess some of the transform transformative aspects, these are essentially new capabilities we don't have today. It could be something like adding AI into our CRM system. So it knows all of our customers. It might know which customers are similar. It might know, you know, what is the what is the optimal engagement strategy based on how the customer acts or what are the predicted behaviors? Yeah. And I think on the on the on the top right, sort of the new products and services, it could be something like, you know, it's hard to imagine even, but it could be something like the always on product expert that maybe today we don't interface to our clients. Yeah. But maybe we can now make that possible to help solve the most complicated issues. So these are just a sort of few ideas. And I think on the on the right hand side, there are a lot of ideas that are new, but there might also be ideas that we are sort of yet to discover. So new revenue streams we hadn't thought about, new core capabilities. And I think I think just the final thing I'll say about this is we see now also in the BCG study, you mentioned Osmos, a lot of people are sort of sticking to the left hand side of the incremental side. And that's the easiest place to to start. Probably it's it's it's a bit easier to grasp, sort of say, how can we improve internal current operations? Yeah. And I think if you listen to to what Anders was saying, so he was mentioning that, you know, they have installed a new sales model, you know, a year back. And this is a lever that generative AI is a lever for actually implementing that even better and get even more quality out of those those investments. So that's that's on the left side. Definitely. Definitely. So we'll we'll touch a bit more on the left side now. And I'll try just to to describe a little bit more practically how we could sort of how you could approach finding out where should we actually start? And and and it's it's it's fairly simple. If you if you look at it from from a classic funnel perspective, you have a buying process that your your customers are going through. You have your ideally your your sales process aligned to that. And whether that is a complex, very complex sales process with multiple stakeholders involved or very or it's very simple and transactional, it's basically the same logic that you would go through. So so start there and then do your analysis, basically. So try to understand where do we have pains or issues or potentials or opportunities within the different stages where we simply need to do better. And and when we and we sort of find out where where is the biggest gap, so to say, it would be a matter of of trying to understand the reasoning underneath and so also to do the proper analysis. And this example here where it's it's just a simplistic view, but start here. That's where we have the scenario of a sales organization that is perhaps too reactive and is is not engaging early enough in the in the customers buying process. Maybe they're not engaging high enough in the organization talking to the right stakeholders. So there might be a potential for for doing something here. What the solution is or could be. I'll come back to that. But but what you can see at the bottom of the slide here is that that there would be different ideas that solves different issues. So are we at at at more on the proposal part of it? Then there might be a tool or a gen AI solution that that helps on generating proposals faster or better. And or as you see now on the on the on the on the on the line where it says start here that there could be a coach that would support salespeople to to prepare even better for the early engagement. So so these are just simple solutions that you could look into. And when you unfold these solutions, you can also label them as as teammates. And I think really this they should more or less be seen as as as members of the team. They may not be on the payroll exactly as the as the rest, but but they are really a great help to help in different key activities that we have outlined already on the formula with with, you know, we can do something faster and we can do things with better quality. These are just five examples basically that helps solving a specific issue. And I think it's it's important to say that these are not, you know, the only ones and it's not like everyone should have one of each. I think it's a matter of finding out where's our need and then identify what would be the right teammates to to help me and help us improve our our both productivity and quality related to this. What's cool about these teammates and we'll show it also in the demo is that that those of you who has been playing around with just, you know, chat GPT or similar know that it requires a bit of prompting. So this process where you basically engage with the language model and sometimes you get good outputs, sometimes not. And I think what we've discovered through our project series that the more we can make these prompts ready made or pre cooked and then sort of put them into the model. So the users don't have to basically prompt, they just have to fill in information or even get the information from a system. Then you get really high, high outcomes of this because you sort of remove the entire resistance part of doing this. So for us, it's a it's a matter of creativity related to these and your needs. So are you seeing a high intake of new employees? Well, then maybe you need a teammate that is an onboarding colleague that knows all about the basics in the company, in the sales, in the way we do sales, etc. So it really depends on the on the situation. And that might also help sort of narrow in the problem space or the playing field, right? Because like don't start with everything at once, but maybe take a look at your paints first of all, right? Yeah, exactly. I think now we've covered sort of the business side. Let's talk a little bit about the tech. Yeah, you can say a bit about that first at least. So I think moving closer to some of the tech, we see sort of generally speaking that at least from our experiences so far, is that the tech sort of falls into three categories. So the first category, what we see out there is a lot of people are experimenting and piloting with open source models. We've seen some do actually brilliant pilots with the free chat GPT version. And that's okay for starters. That's okay. I would say be careful what data you add. It might travel outside Europe. It might be logged on a server in the US. And again, sort of depending on the size and the scale of the pilot, you might need to invest $20 a month license or so. And I think that's okay in the scale of onboarding maybe 10 salespeople and trying it out. If you're trying to move that to 1000 people, I would say go other directions. But again, it comes back to your point from before, which is it requires good prompting. And that is sort of a science on its own. So that's something to note. We're also seeing a lot of things happening with own solutions. Here I put the AI teammates, but it could be a lot of different things. We'll show you one AI teammate a bit later. Other solutions could be reading long proposals or checking RFP for content. But generally speaking, these are great for early pilots and testing. They are low cost. We can build in the prompts sort of underneath the hood. It makes it easy to interface with. But obviously it would require some technical competency setting that up. And then I think third and finally what we see is that we are seeing a lot of things happening in the sort of co-piloting space. So we're seeing new tech come out from Salesforce. They made Einstein GPT. We're seeing co-pilot for sales. We're seeing co-pilot for Dynamics 365. And these are really good. But they are also, they have a bit of a higher barrier to entry because they are a little bit expensive. Sort of $50 a month licenses. And some of them at least without going into too many details are not necessarily fully mature just yet. But again, they hold the sort of premise that prompts are built into the system. They're relatively easy to interface with. Maybe you want to do some training as well. But at least one thing we know is that it does require good data. Yeah. So a co-pilot on a poorly structured CRM or ERP data set will not be a joy. No. So I guess this is just sort of to say a little bit about the tech we see. Nothing is necessarily right or wrong. I think it can be a good idea to try and pick up those early experiences. And maybe at the same time you can clean up your CRM and ERP data. We can at least say that when we look at it from our perspective and those clients we've been helping, it's been mainly on the two first blocks so far. It's been so far. Yeah. Exactly. Let's do a check in. Let's do a check in. Yes. Good idea. So just like in the very beginning, now we've spoken a bit about sort of the possibilities and the considerations of applying generative AI in B2B sales. Yeah. But we're curious to open up the voting floor once more just to get a sense of where you all are today. Yeah. So are we sort of looking at people, have you not started yet? Are you looking for inspiration? Have you begun some testing with open source tools or maybe an internal solution? Or have you sort of gone the full length and acquired co-pilot technology? So we're a bit curious as to how that might look. Yeah. Let's give it a bit of time. There is a small delay, I understand. Yeah, it's a small delay. Yeah. I can say a bit so far. I mean, for our sake in implement, we're actually trying a little bit, you know, building some use cases, building some pilots, some AI teammates, as you call them. Yeah. Just to see how might that help us a little bit. So that's some of what we are experiencing with right now. Yeah. And I think I was talking to one of our colleagues and, you know, he's explaining that right now it's also a bit of, you call it build and burn, which is a new term for me. I'm not in IT space, but we need to learn something here. So there might be investment in building something that we might burn again and then go on another solution. But at least it has moved our ability to make the proper decisions with this regard. Yeah. All right. So we're seeing some. We're actually, so far, we're seeing sort of 40%-ish are testing with open access tools. Yeah. Not a lot of people have not started and 30 % or so are actively looking for inspiration. But we also see some 7, 8, 9% that are sort of have an internal solution and maybe are sort of experimenting with use cases and pilots so far. Yeah. That's super cool. That's cool. Yeah. Okay. Thank you for responding and giving a few insights on that. That leads us to the next topic here, which is, you know, let's be concrete. Let's try to be a bit more practical. Let's hear a little bit more about what they're experiencing in Sanystol and with Anders. But let's also show a demo on how could a team look like just to give a flavor of what it is that we're talking about. All right. So, essentially here, this is just, you know, an overview of the case that you heard Anna speak about a little bit earlier. So, there is some reasoning on the left side on why they're doing it. But if I had to just give a few words on the solution on the right side in this triangle here between a sales leader, a sales rep, and then having a teammate of an AI sales coach. It's essentially not at all substituting anyone. So, it's not removing the leader or removing the sales. It's only a support. So, it's a support into the actual coaching setup, as he was mentioning. So, if you want to discuss a given opportunity or a given account using a coaching framework, the coach is providing questions and support to actually do that and inspire the leader to think differently and come up with new activities. But at the same time, you know, it's also a tool for the sales person alone. So, when there's not a planned coaching session about a given opportunity, this is the opportunity to do self-coaching basically together with the teammates and get inspiration for how to think differently. So, now, we're going to step into what I would call full risk mode. Yes. At least we are almost done. We have ten minutes left, so it's not the worst thing. But I'm assuming it's going to be just fine. So, now Tobias will just take us through an example of a quick demo of what a teammate could look like. Yes. And I would say, first, live demos are always risky, so it might go wrong, but bear with us. And I would say, also, just to show, this is a really quick demo that we made just for this webinar just to give you an idea of how this might look like in practice. So, I'm just writing some stuff here. Maybe you can say a bit, Rosbeth. Yeah, I was actually just amazed about the fact that you could stand and talk and write at the same time. So, I can take over. So, now, Tobias is providing a little bit of information. He's not prompting, I would say. He's just filling in information. Context. Context is probably the right word. Thank you for that. And then, the actual prompts, the requests, the needs that we would like to get out of the model is built in. And you already on the top here see you have different examples that you would be able to ask for. For now, it's a meeting preparation and if you read the text, I don't know if you can follow both me talking and reading the text as audience, but over to you. I can. I'll just say this would be an example of prompts that are sort of baked into a system. Yeah. So, I'm not prompting here. Users would not have to prompt. They would just have to write a little bit of context. So, here I'm just saying something about an old friend, John Johnson, about a sales opportunity and I will have a coffee meeting with him. Yeah. So, I'm going to do a few things just for small context and what we set up here just to show you what this might look like. That sort of meeting preparation in the bottom, sort of top left of the radar from before. Yeah. Here, it gives me some open questions to bring to the meeting. Yeah. This might be one place of starting meeting preparation. Yeah. And obviously, these would then change based on what you provide of the context. Exactly. Yeah. Exactly. So, in the pre-baked part, I wrote a lot about implement, what we do, who we are. Yeah. Now, it knows that the potential client here is called John. It will now help me sort of better grasp the opportunity with John Johnson. Yeah. We can make a lot of things as well. It will also, and you'll see it do with this live, give us an idea about sort of what could a meeting agenda look like. It will maybe generate in the next part an email that I could send to John. Yeah. Or at least an 80% finished email. I have to make some revisions. Yeah. I think it's just worthwhile saying that if you, for example, have this, you want to make better meeting preparation, you would probably have to identify what is the type of meeting that you're meeting with the client. So, is it an early stage inspiration meeting or are we closer to a final negotiation meeting? That would require different prompting underneath the hood. Yes. Yeah, exactly. And I think for now I just made one text box up here, but we could have made buttons or dropdowns for type of meeting and so on, right? Yeah. Then I think just really quickly, this is one way of doing it. Yeah. Another way of doing it would be to make integrations to your systems. Yeah. This is maybe going from pilot to pilot plus. But instead of having to write a lot of context, maybe I can just fit it into my context of I have some leads in my CRM system. Yeah. Maybe I can revise those into a prompt. And then suddenly we have a solution that now this is of course just sample data. Sure. But essentially allows me the same without all that preparation. Yeah. Now I think this is really cool because this would also do some open questioning. It would give me some potential objections that the client may have. And it will also help me type into the implement network. So it will say, okay, this opportunity is about a certain topic. Maybe you should refer to your colleague in another practice in implement. Yeah. So maybe if I just can add to that. So if you remember the stats from the BCG study, they said that, you know, there's a potential of uplifting in your cross sales in an organization like implement and also many other organizations, you know, it's sometimes hard to find out how do we, you know, take along with other applications that the one that I'm as a salesperson is maybe the expert within. So here we get the opportunity. I don't know if it's working that fast, but we would be able to find who would be the person to contact to basically join a joint sales call, for example. Yeah. Exactly. So it reduces some of that friction of actually just getting the internal network in the organization up and running. Exactly. And I think the final point, if I may, you spoke a lot about coaching also, Rasmus. And I think one thing here that makes this also cool for meeting preparation would also be something like this. You know, these would be some objections that John may have. Yeah. These may be some strategies to overcome. And these are some coaching tips for me as a salesperson to sort of note and bring forward. Yeah. We would love to show way more about this. I don't think we have the time. Probably not. So we'll probably cut it short here. Yeah. But this is what a pilot could look like. Yeah. Just to give an idea. Yeah. Okay. All right. Cool. For the last five minutes, then we are on the clock. We'll just share a few insights, some perspectives, and perhaps a little bit of advice in terms of moving forward. And then we'll wrap it up for today. Initially, here's just a few insights from another project, which is anonymized. But we did something where we actually focused on the meeting preparation. And what we saw on the left side on the impact is that it actually boosted the speed of preparing. It improved the quality and it led to better meeting, customer meetings and conversations. And I think what's maybe what stands out for me is that some of the users were saying that in a super busy day as a salesperson where I have multiple tasks that I need to handle, just the fact that I had to spend maybe just like five, ten minutes in actually sitting down and preparing for a specific meeting was super effective. So it just helped me on the focus side. And I knew I sort of didn't have to go into a half an hour, one hour task where I would be interrupted over and over. So I think that's really interesting. And then I think the next part is that how do you make that sustainable in some way? And that relates to how we embed leadership. So I think that's an old saying in general, but also in sales. And it especially is also here. So leadership reinforcement is super critical. So in the beginning here with the old example from this case here, we had leaders simply taking one to one conversations and taking in with this meeting preparation, how is it working to show the attention and the interest within this to make sure that this is a part of how we want to do and how we want to operate to actually sustain that change. So think of leadership as the core still on how to make this happen. And this is just as any other system, whether it's CIM or whatnot, you know, it needs to be reinforced. And that speaks to the adoption and change part. A few final reflections? Yes. Okay. Absolutely. And maybe some mistakes. I will phrase them positively instead. Do that. But I think, you know, we have done a few of these projects already, Rasmus. And we know both from these types of projects, but also from others, sometimes pilots, projects, products, they go to die. They are no longer relevant. And I think to sort of grasp that in a generative AI sense, one thing I would say is early on, don't make it an IT project. If you're going for production, absolutely bring in IT. We're not saying to keep them out, but at least in the early stages, don't make it an IT project. I would say, secondly, try to bring on expert prompt engineers. These will allow the systems to give the best possible outputs, and they can take the prompting task out of the hands of people, so they only have to do context. So writing like I did before or choosing from a dropdown. And then I think finally, don't forget the change communication around this. And as you alluded to, leaderships can reinforce this being a positive change in sales organizations. So I think that's sort of the main point from this slide. Yep. All right. We are running on the clock, Rasmus. And I think from our side, before we let you go, before we say thank you, we would love to hear for this short morning webinar, what was the best learnings or what are your sort of main takeaways from today's session? Please write one point in the chat, or maybe two if you have it. And then we're curious to follow along. Yeah. So I think the chat is open, and I think I'll just use the time while you respond to say thank you for checking in. It's been fun preparing. We didn't spend three months, as I said in the beginning, to actually get to this date. We waited a little bit, and the learnings, they keep on coming the more we work with this. And it's super interesting to work with. And now we're also seeing the chat coming alive here again. So while that is coming alive, I'm just going to say thank you so much for joining. We're happy to see that so many actually want to spend the time here this morning. We know you have busy schedules, all of you. And if you have any questions, any need for a conversation of any kind, we are always ready to talk. So please reach out at any time. Anything from YouTube? Thank you so much. Yeah. Have a good day. Bye bye. Bye. Bye bye.