Building a Practical AI-First Sales Organisation

0:06Casper Guldager:Welcome back to the Mastering Sales and Negotiations podcast. I'm your host, Rachel Massey. And this week on the podcast, we were joined by Casper Guldager, AI strategist, former Lego and KPMG leader, and someone who spent the last decade helping organizations actually make AI work. Casper cut through the hype around AI in sales and shared what really drives adoption. From leadership behavior and daily discipline to why most salespeople give up on AI far too soon. We explored where AI genuinely adds value, where human skills still matter most, and how structured approaches like spin help salespeople get far more from the technology.

0:48Casper Guldager:If you're trying to turn AI from an experiment into real commercial impact, then this one's for you. So let's dive in.

1:02Casper Guldager:Kasper, welcome to the show. Thank you. So it looks like you spent the last 10 years working in AI consulting, so from Lego to KPMG, and you're now running your own venture. So can you talk us through that journey and what led you to focus on sort of AI adoption rather than just the technology? Yeah, so I can actually start maybe 18 years ago, I actually started in telemarketing. And that was actually one of my favorite jobs ever in working in selling papers, newspapers in the city I was from, from Aarhus. And I super loved that. And actually, already back then, I started looking into practicing my speaking or my selling pitch to just simple voice and listening to it again and again.

1:57But at least there was a super interest in using AI in the way I worked. And then, yeah, I started in Lego when I was midway in my university. and I actually think one of the biggest learnings there was that technology worked but people worked against it or were often saying yeah but I could have done this small part better but it would have taken me a month and stuff like really challenging and and because if there was just small stuff that the technology was not doing and I think that just catch my interest and actually it's something I'm reflecting about later in my career it was not something I reflected about back then okay now I'll just do adoption the rest of my career I just it just happened because there was so much resistance from people every time I worked with technology and that was

2:58Casper Guldager:you mean that people weren't seeing the bigger picture with the technology they were just going it won't do this and they weren't seeing that this is also what i could achieve by using it yeah that and they often also just argued against using it because they were like yeah yeah but i could have done this small part better yes congratulations dude it's uh but what about the 10 other things that the technology really delivered well on it it was it it was a super interesting but i was also super young so i was also easier to push away from what was the target but lego lovely place and you get to play with toys every day it's crazy um but in anyway there There was a lot of my friends moving to Copenhagen, and I also wanted to move.

3:57And then I started applying for some management consulting roles. And I came into KPMG and was super, super happy of joining their graduate program. And there I was exposed for their team that was called New Tech. And I worked on their projects. and there was again you know we were building and here we were even building more stuff for clients and they had paid us to come as consultants and i could again see there was a bit like it's not so i think just my whole career i've experienced and we also really enjoyed that interaction and teaching these people how to actually work the best way together with artificial intelligence and then you can say the last three years have just been crazy and everyone have had them in their hands and i've been so blessed and super lucky that i was sitting you a place where I've been working with this topic for so long.

5:12So I've just been so happy and been invited for conferences around the world and worked with the most interesting problems and clients about adopting all these new technologies, adopting TGPT, Copilot, Claude. How do you adopt all of them on the same time? Is some of the private equity companies asking me. You know, and that, wow, it's just, yeah, so fun.

5:41Casper Guldager:And you argue that sort of now it's the time for AI to be really embedded into sales teams. Why do you think it's so relevant now? I think it's, it has always been relevant, but I also think at least I've been involved in quite a lot of commercial teams and commercial projects. And often sales are also staved by all the traditional automation because it's like, leave us alone, give us a CRM system, and then let's go out there and sell stuff, which have been fine. But now it's so easy to try. I've seen some of the commercial teams I have been teaching, they are actually really having super innovative thoughts about how to use these tools.

6:32I have seen commercial teams use it for route planning. You know, okay, I'm going, I want to spend more time with physical time with my clients, giving them AI their calendar and asking it to help them reorganize and use more physical time with their clients. You know, it was not something I have thought about. I'm really seeing commercial teams having a huge opportunity because it's so easy. to use these tools towards what they are using. And then I also think if you're actually looking at what does commercial teams do, they have an interaction with a lot of people and try to align, at least in advanced sales, and try to align an incentive.

7:23AI is completely magical to do that because you can prepare better, You can actually keep notes of all the meetings, ask what did we discuss in the last three meetings, what is the incentive. So, yeah, I think there's just a million arguments for why commercial teams should go full on adoption into these tools. And Casper, do you think that that adoption, you mentioned earlier that idea that people were resistant because they were looking at the bits that it didn't do, rather than the bits that it did do, which were in many times, you know, many more actions that it could do really well really quickly.

8:06But we focus on that little bit where things weren't quite perfect. Do you think people have now got over that because they've been exposed to, you know, ChatGPT or Copilot or whatever it is that their organizations have got and they've realized that it's not a case of pressing a button and getting a response anymore, but there is an interaction process which means that it gets better and better and better. Do you think that's helped the adoption, but has it also given them almost a kind of creative spark to say, oh, I wonder if, and how has that played out? I think I will answer the way of, there is still a lot of doubt from commercial teams towards these models because most actually have tried to ask their company chat gpt or their company own build chat and asked it for something advanced regarding their product and it answered wrong yeah because it didn't get the right articles or they didn't get the right references or so it and then they are so many have come and showed me look look look look it's it's answering wrong and it's hallucination and sometimes it's simply also just because it have been built not the best way their internal technology because it have been early days but that is their experience and i completely respect their experience so so i'm normally telling them it's the same if if you asked uh one of your colleagues and they answered wrong it's not because they will answer wrong the rest of their life it it it no but it's it no but it often isn't right and but they did it but we have a tendency with technology to to really give it a hat on if it was wrong once wrong forever now um so so there is something there and i think for all the stuff that really can say works for them right now or motivate is actually that they can see other commercial personnel totally killing it yeah you know there is though that i have met them around the world at conferences also at at customers and and big organizations where there is just that you know commercial rock star that have agents sitting and searching for all the new information on all their accounts they are sitting and preparing for each meeting they get a script and a small podcast they can listen to so they're way better prepared for meetings you know they're they've just set up a full a system of a lot of agents that does help them power through doing sales and when it's not me telling when it's you know their colleague and they are just oh going there that that's inspiring and that makes other people adopt sure so i think there is those two forces happening right now there is the the strong colleague force and i have asked ai back then and it was wrong and then those two are battling it yeah yeah it's a bit of tension there isn't there so you you mentioned there that idea of being able to um essentially use ai as part of that preparation piece that that getting information getting data understanding the recent changes within either a client or a prospective customer's perspective what how is that feeding into what people do face to face because you said they're also using it to maximize or at least optimize the amount of time they speak with their clients so so how how are those two things being pulled together by those organizations who you described as rock stars?

12:04Yeah, at least what I have seen in some of these commercial teams where they are really rock starring out there, they have used AI simply to take pictures of their calendar and said to AI, hey, I would like to optimize towards being more physically at these clients Right now, they are prioritized for me, and I want to spend less time on admin work for XYZ proposals, looking into their products, understanding their stakeholder landscape, understanding my own internal landscape. That can also be a big thing in bigger organizations. and then start digging down and saying, okay, how can I do AI for, how can I set up some simple workflows or so simple, it can be as simple as a custom GPT.

13:01I have one software company I'm helping right now. They have a custom GPT. They call it a stakeholder GPT. It's a small solution you can build inside ChatGPT. You give it instructions. And in this case, they said to this stakeholder GPT, you are a stakeholder GPT. You are representing all our 15 key stakeholders on this account. Then they did deep research on all those 15 key stakeholders who was on their account, fed that data into the model. And now they actually have a place where they can simulate meetings. They can ask it for, okay, what would the incentive if we are trying to bring in these new parts of our product, what would be the people that would say against it?

13:54Who will be maybe a bit for our suggestions and how do we manage that? So really simulating their account. And it's a super simple technology solution. Everyone can set it up. it's six seven steps but it's super powerful for them as a group um so yeah i think those rock stars that does it inspire the others because it's not because they they have spent so much time

14:26Casper Guldager:they yeah so it looks like then in that in that case that they've taken some they've taken their technology and it's not ai as a technology tool it's almost moving into that being a behavioral tool because you're actually using it to get better at something that you need to do as a human precise actually i love that you're saying that because i really reflect a lot about because a lot of people also ask me what is the value of ai and i'm often asking them what are you doing that is valuable today okay let's try to figure out how ai can help you do that more better faster all because that is often the answer so yeah yeah it gets down to some fairly basic things doesn't it you know one of the things that that i talk about a lot with with with our clients is quality quantity cost or time it's like where are you moving those four levers and if you can do it better you do more of it in the same resource if you can reduce the cost if you can do it quicker then you're getting benefit in some ways and and and actually you can work out that value in fiscal terms if you want to you know not that most people need to go that far but fundamentally there's a way of making it more efficient to enable you to have more time doing the things that really count you mentioned that in that stakeholder tool that the the team themselves have built it is is that is that common now is that is that what you're seeing it's not giving it out to the it team or bringing in some expensive AI company that will build it for you or whatever it is.

16:04Is it really that simple and that easy to use that people are building their own agents to do their own work? I love the question because that question also evolves every day. Because right now, if you have thought the last two weeks, it's going crazy what you can do with building tools with AI, cloud code and cloud cowork have exploded lovable have exploded and like there's such a big movement within this space and my answer here will be that all the teams i work with they can build some stuff that can really really help them themselves yeah they should not wait for it You should not wait for anyone.

16:52They should just do it themselves. Prove that there is a business. Then some teams, when they get more mature, they're saying, they're coming to me and asking, okay, this stakeholder, but could it actually also update and look in the internet and look for some trends on how is it going on that account, bringing it in? Could it actually also remember all our previous talks with all these stakeholders and have like an integration between, for example, our teams where we record all our meetings? And then I'm saying that is something that needs to be built a bit more. But you have proven, if you have built it yourself the first, you have proven the business, you have proven the value.

17:45And therefore, you can discuss if you should even build something. Because I truly believe we are living in an age where you can't come with a specification for a system you would like on a Word document or a PowerPoint. you come with a working system that you want more features to. Yeah, that's really interesting. That idea of kind of incrementally building on what you've got, but starting off not with an idea or a piece of paper or a specification written on an Excel spreadsheet, but actually saying, look, I've built it. It does this. It'd be great if it did X, Y, or Z as well. You know, it's that sort of idea.

18:32And it's precisely that. And that is also what I see. The most mature sales organizations I'm working with, they work like that. They have tested all these small ideas themselves, and they have had all those 10, 20, 50 iterations making it better themselves. and then in the end if they are you know they want integrations they want something that is simply technical can't be done in the technology they have access to then they discuss and it's again it's also easier to handle for someone that needs to help them build it if they have all that knowledge and work with it for some time.

19:12Casper Guldager:I think one of the things that I understand your belief is that it's up to the sales leadership. They are fundamentally responsible, really, for ensuring that AI is adopted. So in these sort of successful sales organizations that you're working with, is that what you see? Yes, that is my deep, deep say to the world. And also one of the reasons why I went out and started a company that is called Executive AI that is actually now part of a company that is called Adaptic. And what we do in those companies is we teach leaders how to be a role model. We teach them how to use these tools deeply in the way they self-work, both in their more commercial role as a sales leader, but also as a leader.

20:12How can you use Copilot or whatever tool you have internally to make searches for all your people and what is the interactions you have had with your people? Get that summarized before you have one-to-ones. So you're way more practical and pragmatic when you have interacting with your, yeah, you're actually the people you lead. Because if you are a strong role model and you really use these tools, then there is a way bigger chance that that trickle-down effect in your company will happen. And you are also, as a leader, you're simply responsible to use this to turn the key towards something that is valuable for the overall organization.

20:57Do we want to use it to have better customer satisfaction? Do we want to sell more? Or do we simply just use it to have better terms for our commercial teams so everyone wants to work for us? Or is it a combination? But that is you as a leader that needs to figure that out. And therefore, you should be deep, deep into these tools. And that is the only thing I'm working with right now is bringing these AIs so close to the leadership as possible. I think that's really interesting because we teach coaches, we train coaches to work with our models, Spin in particular. And when we do that, we say, look, if you want people to prepare for their meetings with their customers or their prospective customers, but you don't prepare for a coaching conversation, what messages are you sending?

21:52And now you're saying, well, add on to that. How are they preparing? What tools are they using? What is AI doing for them to speed up, to deepen the information that they're getting? And how does that better prepare people for the conversations that they need to have in that kind of chain, if you will, of conversations through the organization, which is really interesting? Are you having any of that resistance to that adoption from those leaders? do they get it? And there's a kind of an experience that we've had, I've had particularly in talking to people, which is that sales organizations have had an AI tool imposed on them.

22:36So it's come down from the leadership sort of, and it's almost always driven by can we achieve more with less? So there's a cost management side to it somewhere down the line. has that now gone or is that now going? And are leaders really embracing the idea of, well, look, I've built this tool which helps me do my bit of the job. And I strongly recommend and can act and support you in building a tool which will help you do your bit of the job more effectively and more efficiently. So if you look at a lot of the big surveys out there, there is a higher adoption in the leadership than there is in the general public of and so that is a fact but if if you look from a practical point of view where where i'm coming from and with all the actually the leadership teams i'm working with they they have used it and they understand it on some level because they have had this discussion the last three years about what should we do about these tools.

23:53And often it had been a checkmark thing. It had been, let's buy a system and then push it towards the organization and then say, let's go. That was done. Or they built some internal tool and then pushed it and then said checkmark. But this is an ongoing process And sorry to say, but I think it's an ongoing process for many, many years from now because everything just changed every year. So, I think that leaders should be way more active in using these tools on a general basis, working with it, and using themselves as role models to show and to demo towards the rest of the organizations to create the trickle-down effects.

24:43A lot of people still sit and think, am I allowed to use it? Yeah, yeah, yeah. And having that stamp from the leadership that I'm using it and I'm using it a lot just creates a lot of positive effects in a company. And then I think because your question was also a bit towards what is it companies want to use AI for? And, of course, nearly all they always say, okay, let's do more with less, as you said. But I have a feeling that more companies also start thinking about that AI can maybe completely change the way the company delivers value. so it's simply just that changing their business model or completely changing the way they work within different areas of their organization and that is a way bigger leap and of course also means way bigger work and um yeah and the ones that does it maybe they will just be 10 15 x the others and then they will be super hard to catch yeah yeah and that's really interesting what what what do you think and just thinking about that from a particularly from a commercial perspective what's what's the really kind of astonishing or or brilliant examples that you've seen with those organizations where they've taken that trickle-down approach they've got people using ai to solve problems what what give me an example of something that people have done where you go okay that's really good you know i've seen a lot of ai that's a really good example what's the sort of stuff that people have been able to do?

26:34I think what companies, at least, you know, what they say, this is great, is actually often okay, simple, and that is often how it is with stuff that works. But I think it is a leadership that actually really have shown, okay, we are using this to understand our competition. we're getting newsletters on our competition generated by ai every month that compares with our own strategy it looks into our sales numbers and then it digs down into all the external numbers they can find and then it gives some signals on it so it gives like all those decision points that people can use, but they are just strong and they are consistent in something that happens every day over a very long time, which means that the people get to trust the system and they are willing to act based on it.

27:38I think that is on the leadership. I think on the process side, I think actually some of the examples I gave before, some of the most like that have really created some super impact have been the companies that have been willing or the commercial teams that have been willing to nerd their customers way more and what i'm meaning like nerd their customer people like really really you know digging down in all news about their customer, all news about the people that is in that organization, and really, really have a deeper conversation with AIs on how can our product help this customer win more in their situation?

28:33Because that is what it's all about in commercial and in sales. It is if you can make your clients win then they would often like to buy more yeah so yeah i think that's interesting

28:47Casper Guldager:because i think people tend to think of ai and the research as this big picture thing but what you're saying is you know get really deep into the detail get it very very specific yeah i'm saying that and then i'm also saying that the companies that succeed they are actually not it's not the use case they succeed with it's the system it's the system of that the leaders use it and they are good at showcasing the rest of the organization that is that the commercial teams and actually also all the other teams are willing to build these custom gbts or small solutions that helps them and test and business validate all the ideas they have and only scale the ones that really can deliver big value.

29:34Because it's the system that makes you successful. It's not the one solution. Because everyone can build a good solution if they just pay enough. But making your company really structured, the way they actually work around AI is the big thing. So there's an ecosystem at work there. And I really like the idea of the cadence of use builds both opportunity and capability. So there's a trust thing. There's a kind of, okay, I get this now. But there's also, oh, it could also do this. And because you're immersed in the ecosystem, it opens all of those possibilities and gets people thinking in a different way about what they and the systems that they have around than might be capable of.

30:28I think that's fascinating. If you were to advise an organization which is currently sort of interested in, but they've not really got engaged with AI yet, they've not done a lot, maybe they've got a corporate license to Microsoft Copilot and permission to use, but that's as far as they've gone, okay? So what would you advise them to start with if they wanted to start getting some real value from using what is quite a transformational tool? I often use a simple target, and that is 100-10-10, and that is 100 % adoption on your leadership. So they should all use the tools and be able to explain and be role models.

31:20You need 10 % of your organizations to be the people that are AI champions, or I'm actually liking the word AI influencers, because they should be both great at using the tool, and they should be great at telling and explaining and showcasing the rest of the organization how they gain value out of it. And the last 10 should be AI builders. That is the people that can build the custom GPT, that can spin up a small agent that helps a team, And if you have 110.10, you will see a lot of magical stuff happening in your company. And then you can say, how do you start there? You simply start by saying, okay, how do we get the 100 % adoption on the leadership?

32:14It's often something with KPIs and offering training. the 10 champions it's a bit of a tougher work because it's creating a champions networks that that really work and and they are actually willing to keep teaching the rest of the organization it's a lot about setting up the right communities giving them it but it's yeah you have not heard me say anything about technology because all this can be done in copilot all of and and if you're maturing a bit more up as an organization and your organization scream for having access to more tools maybe just give it it's it's not a a but let the organization prove the value with the 11010 system that i would recommend that's really so just just for sort of information you're touching

33:11Casper Guldager:there on what what um is sort of known as a agentic ai can you explain what you what will you what we mean by that yeah of course so so i think that it's actually very simple because you could go very technically into this but there is the llms until recently what they did for you was that they gave you advice so if you ask chat gpt you they gave you advice to what to do. They gave you advice to how to interact with your clients. They gave you advice on how to write a good email or a good outreach to your customers. What the agentic AIs do now, they can act. They can do something for you. They can go out and troll a lot of different websites trying to find leads.

34:06They can put them into an Excel sheet. They can start booking it in your calendar to look into it so they can act the base on you. And you can also ask many agents to work together on a bigger task. And then they, as long as you are good enough to give them an end goal. And that is the hard part now is, but it's again, what have organizations done for the last many, many years? They have made people work together towards some goals. Now we need to make people and agents work together towards some goals. So Agentic is doing, and all the other models is advising and giving you good ideas or answers.

34:56Casper Guldager:I mean, if you think about what that's doing now, and if I think about what I was doing maybe just two years ago, I mean, it's so different now. Where do you think salespeople or commercial organizations, what will they be? Three to five years' time now, what does it look like? When I reflected about this, because I get the question quite often, and I think it's a lot about when you look at your calendar. I look at my calendar every week, and I have a lot of different AIs that challenge my calendar, that say, okay, how should you use your time and stuff like that. I'm very aware of how I use my time.

35:37And I think in the future, we will spend more time on stuff that we find interesting. So in three to five years, a commercial or sales role, if they look at their calendar today and they took a pen and said, OK, what I like the most here, then they would do more of that and the stuff they like less they would they would do less of because that is how actually at least most people i've seen work with ai they are good at amplifying nerding doing more of what they like and where they are really or where they are good as persons and they they try to delegate and move away the stuff that they are not good at.

36:30Casper Guldager:Yeah. So more chatting to customers, less complete in your CRM. That is a different way of saying it. But, you know, from my time back in KPMG, I also remember, you know, that there was quite a lot of time on not being together with the client. And I for sure, I think less CRM, more client time. Yeah, absolutely. Absolutely. Absolutely. If you were to think back, when you first got involved as a junior consultant within New Tech and KPMG, if you were speaking to somebody who's in that position now, within either a consultancy organization or within an organization, but they got access to some of this kit, what would you say, do this?

37:21This is the thing that you should do now, which will help you in the future. What would that be? I think regarding what everyone should do now is getting so familiar with these tools and figure out how they can deliver value with them and how they love to deliver value with them because it's so different. I've seen people that are super great graphical and like that mindset of how things should look and like look and feel and branding and stuff like that. they can be amplified in that and people that are super great at building like code and stuff like they can be super great at that so i think a lot of us should reflect more about what is it that we really really love and enjoy yeah and we have maybe been told for many years that you also just need to make a lot of money and have a dog and some kids and and that is also true but it's like what is it that makes you valuable and then and then there is and the reason why i stopped in kpmg and went out and and did my own is because i'm totally out now and using all my time all my time i can on teaching because teaching there will be so many years where this will be a skill to learn.