CTRL/SHIFT - 14 06 2026 PARTHENOPE STAGE - MAURO LUPI - AI LEADERSHIP SKILLS

Good evening, good afternoon. So, um, just to break the ice for a second, how many of you use any kind of generative AI application on a daily basis? Ah, not all of them. Okay? I was really imagining something global.

Oh, a few words, and then they'll obviously make these slides available to you. I also deal with artificial intelligence. I teach this here, adoption of AI in companies at the University of Milan, but in reality I am the manager of a business team for Serchon, you mentioned it before, a web agency, but quite historic, let's say. I'm a bit of a digital veteran. Well, what I wanted to talk to you about today is a bit like this .

So, now I understand how it works. So, I think we all now have at our disposal a very powerful, new tool, a new technology that is actually very different. It's not the new Excel, I think we've all learned it a little, right? We have something in front of us and in our hands that is up to us to know how to use, right? Unlike now, you use Excel, okay?

Do you know him? Someone? Yeah, okay, but I mean any classic instrument, if you think about it for a second, is pretty much the same for all users, right? Then, clearly, it's up to us to know how to use it more or less. Well, EI isn't like that.

eh you don't have a manual, we can't, there is no manual for cici GPT or copilot, Gemin or whatever , it's up to us to find ways to use it. Now there's a metaphor that I quite like and that has also been taken up a bit by many analysts, by many actors. Even we sometimes, when we integrate and adopt AI in organizations, we think about this metaphorical idea, that E is a bit like a new colleague of ours, okay? Well, then we'll see that it's not to humanize you, absolutely not, in fact this anthropomorphization is conceptually wrong, but anyway, follow me on this metaphor for a moment . Um, a new colleague who has many characteristics that we sometimes also find in colleagues we actually have in the company, right?

Sometimes not, we have a know-it-all in the company , right? The person who knows everything, right? that whatever we ask him he knows anyway. Or a person who is very servile, right? He's always willing to help others, maybe even anticipating their needs, right?

Eh, or the people who really know how to do everything, right? They manage to, um, find their way in any type of context, any type of process, and then there are the people who take a break from their work, the ones who I don't know how they do it, but they're tireless, they work incessantly, they don't seem to have any problems with fatigue or anything else. Well, here you go, you obviously have these things typically, he knows how to do them all, right? It represents them all a little bit. Well, and in this metaphor where conceptually, whether they are agents or whether it's simply our nice chat on Geminai or whatever , we can for a moment consider him our new colleague.

Now, what's the point? There is also this logic. We have someone who helps us get things done. What's important, in my opinion, is that in this logic, this guy here is the head of Salesforce, the owner, the founder of Sales Force, one of the rich guys in Silicon Valley, who says that more and more, whether it's me or whoever, whoever runs a company, but I say in general, whoever manages a team will have to address this issue of having a physical team, a Human Force combined with a digital, virtual team, call it whatever you like, right? Eh, what's the point?

I think we should try to remain the leaders of this team, right? Well, let's consider these tools as new colleagues, shall we? Whether it's the intern who knows little, but actually knows much more than us, or someone with other types of characteristics, the important thing is that we maintain some of this leadership, let's behave like modern leaders. Now the leader, no, some of you or some of you manage teams, but have responsibility for a team of people, even two people, right? Okay, so we're not born, as they say, learned, right?

In this sense, I believe that leadership—that is, the ability to be a manager or, in any case, to manage a team of people or to have responsibility for a team of people—is acquired through experience, through study, through leadership courses. I mean, it's a discipline in its own right, right? you have to know how to do it. Here, with this same logic, I would like to bring you this metaphor because many of the leadership characteristics, which I have happened to teach a little about leadership over the years in my professional life, today we find them exactly the same when we have to manage modern digital tools and in particular those related to artificial intelligence. We must be leaders capable of managing not only humans.

And some of us will do it, some of us, vice versa, won't, maybe they're freelancers or independent or don't manage teams, that's fine, but we still have to manage virtual digital teams and so there are about a dozen of them and I'll quickly go over these leadership characteristics, I find them equally applicable to how it can be useful for us to have as a reference today in managing artificial intelligence. Let's look at them together for a moment. So, one of the classic characteristics of a good leader is knowing how to manage, choose the team, right? Understanding what people I need, what skills I need, seeing what I'm missing, no, in my team based on that, what my goals are to achieve and understanding what I'm missing. Here, in the same way I believe that each of us, I don't know if we do it daily or a little less, I believe we need to ask ourselves a little about but which are the applications that are really useful for what I need for my purpose?

And this is a little different from how we chose technology until yesterday, where there were platforms, no, that more or less had similar features, some had better ones, some worse ones and then we made our comparative choices, we chose Excel rather than something else, right? Google Sheets or I do n't know. Well, today this topic is a little more subtle. You've also heard about, uh, geopolitical scenarios before, right? Well, in my opinion, given what happens every day, we also need to consider these logics and at least be aware that our technological choices regarding AI will affect our professional activity because often these platforms not only constantly change, improve, add functions, but as we saw the day before yesterday, they can suddenly disappear from our hands because a nation decides to change a law that however has a global impact and we find ourselves having perhaps purchased an application that someone says can no longer be exported from the United States and takes it away from under our noses, no, without us being able to decide otherwise.

We must know how to choose, but we must also know how to ask. This is one of the most important characteristics of a good leader, right? to the fact of not being the boss of a team, but having a professional relationship with your colleagues, eh knowing that no one has, how to say, absolute knowledge or the truth holds the truth, no one holds it. It is precisely from the awareness that a team made up of exchanges of mutual enrichment, the fact of knowing how to ask one's colleagues, no, when I want them to perform a certain task, I have to be able to convey it to them, right? If I tell a colleague of mine , "Look, please improve this report or this document, I 'm not telling him anything, I'm not giving him instructions on what to expect."

So, in the same way, we know very well that when I have a GPT on duty and I say to him, "Improve this report or this document," I am not likely to get disastrous results because I have not specified what I want to improve. So knowing how to ask, knowing how to interact with these platforms is our responsibility, right? In 99% of cases, even from practical applications, from the workshops we do, if the result of these platforms doesn't seem adequate to us, even stupid, in the vast majority it's our fault, okay? And that 's a different way of looking at technology, if you think about it, right? We must know how to ask with the famous prompting techniques at all levels.

Well, we have to know how to educate, this is another of the responsibilities of a leader or a team leader, right? The fact of helping your colleagues grow professionally, right? eh to ensure that in a world that is constantly evolving, growing and changing they must be constantly updated, trained on everything that is happening. So, in the same way, we must also make the platforms we use increasingly more suitable, increasingly knowledgeable about what interests us, feeding them with the right data, but good data, because otherwise it's garbage in, garbage out, as they say in IT, right? If I throw garbage in, I will receive garbage.

Um, and so by training these platforms we can also help them make their , let's say, um, results more suited to what we need, right? I'll give you a very practical example now, probably something you all use, but I see that every now and then someone uses this function which is very banal, but I mean it's a really strategic use of chats. I don't know if you use this thing too. Simply put, when I start a new project, acquire a new client, begin a new course, launch a new initiative with my colleagues, etc., etc.

, before opening a folder somewhere in the cloud, etc., we open a chat, no, in our artificial intelligence platform . We typically use Gemini internally or in the Cloud, depending on the type of project. One of the first things, we create an area inside our AI before actually creating a folder where we store the files, because ultimately our work will start from there and everything we create in there will create a shared, collective knowledge base, so to speak, of reference on the topic we are developing. And as you well know, then even after months I return to that area and I don't have to start from scratch, but I develop and increase my knowledge.

Oh, now one of the interesting things, especially for those a little old like me, who typically have to unlearn a lot of things, no, before learning the new. This is a bit of a struggle, sometimes even for those with gray hair, which is to understand what we can actually delegate of our work. Perhaps not delegate entirely, in fact, preferably we don't delegate everything because that would mean we're replaceable, but we try to understand which activities of our work are appropriate, precisely, and where it 's appropriate to get help from these systems. When it comes to integration plans, we do something slightly differently in companies than has always been done with technology. Normally with technology, digitalization, right?

All these processes here started from understanding what technologies do, right? New technology has come out that does this. Great, then let's see what that function can be used for. Now, however, we can almost allow ourselves to do a bit of the opposite and say, "Okay, let's actually analyze everything we do today, that is, let's dismantle our business processes, which ones are the ones that do what, okay? And we can reasonably think of intervening with each of those processes, with the support of AI, sometimes in a marginal way, right?

There are some incremental improvements, something like that, sometimes absolutely significant or even totally disruptive, to say, look at that process there today, let's completely eliminate it and have an agent do it directly, right? In the more advanced cases, of course, but let's start with the processes, not with what the technology does, because otherwise we'll lose focus a bit . One of the words that is used much more often in our sector nowadays is that of the orchestra. Hmm. So, even a manager sometimes has to orchestrate their team, which isn't just making things work, etc.

, but understanding well, precisely, the tools, no, that they have at their disposal. And each tool, as if it were an orchestra, It also has slightly different characteristics, right? Just like the people on the team, no, they have different characteristics, attitudes, er, behaviors . There's the more open person, no, more like that, there's the more conservative one, there's the more resentful person , there's the one who's not, more optimistic, and so on. So, in the same way, fortunately Leai doesn't yet have humanoid characteristics like the ones I mentioned, but undoubtedly the orchestration, the need to make them talk to each other, will become a crucial issue.

It's already like this today and be careful, the illusion that sometimes vendors, no, platforms tell us that everyone talks to everything, but what's the big deal? You go there and connect your calendar, connect Gmail, connect this, connect this. A isn't always easy. B isn't always safe. It's up to us to activate these very useful connections, but in a secure way, so that our data doesn't disappear on the fly or be arbitrarily deleted by a tool.

It's up to us to know how to orchestrate the issue of platforms, let's say, also considering that we now find ourselves You have them in any type of software, not just any hardware connected to the network. Until a few months ago, until last year, at the beginning of last year, I was also teaching courses on refrigerators, for example, which maybe one day, no, they'll have them. In fact, they've been on sale since the end of last year, since Christmas last year. Both Samsung and Samsung make them, and they're basically refrigerators that have cameras inside and, based on what you have in the fridge, they recognize it. They make recipes for you, right?

Or they make a list of products that are about to expire and tell you when they're about to expire. So, this is to say that you have them in the middle of anything connected to the network, and we have to get used to knowing how to orchestrate it. So, another crucial leadership characteristic, a good leader, is also recognized by how well she is able to listen to her colleagues, right? When she's able to interpret their needs, their requirements, er, to perceive their stomach aches , but maybe even their enthusiasms, desires, etc. Now, with this logic, I find it very productive and I see that even today, even our simple chats with Geminai or CHGPT aren't always done this way.

The standard use is to ask something and receive an answer, right? That is, you do this and the application does it for you, rather than using it as a sparing partner, as I say, as a source to question what we do, to have a critical opinion on what we do, and it's extraordinarily useful, okay? You also grow professionally by questioning what you've done until yesterday, right? It's always been this way. Now we have a sufficiently neutral tool, but we simply need to be capable of avoiding the excessive indulgence of these platforms that are programmed to prove us right, to tell us that we're fantastic, that we have brilliant ideas, blah blah blah , right?

They're programmed to prove us right, but it's up to us to find ways with appropriate prompts to ensure that this doesn't happen, but the critical opinion we can get is spectacular, oh very useful, let's question ourselves, it's good for this. Well, clearly our responsibility is to know how to control, right? Because these platforms make mistakes, we know that, right? The theme of hallucinations, as they are generally called. Well, think about it for a moment, but it's the first time I set up a digital agency, I don't know, 40 years ago, I don't know, 35 years ago, it dealt with software, right?

But it's the first time that a technology explicitly says, quite clearly, that it can make mistakes. Mh, you know, right? the little line under the Think about it, it 's never happened, huh? It's not like someone sold us, say, a car, told us, "Look, it usually runs fine, but sometimes it doesn't brake, right?" Or I don't know, someone sold us Excel, right?

And says, "Look, it usually runs fine, but sometimes it gets the sums wrong." Well, but for goodness sake, technologies make mistakes, okay, but they are exceptions, they have their own Bugs, but those are fixed, right? These platforms, on the other hand, systematically make and will continue to make mistakes. Output control, if we simply copy and paste what we see in chats or if we take everything the agents do as true, it's up to us humans—well, we can't delegate. No, the responsibility for these errors lies with us who use them.

So let's not underestimate this ability, which should be one of the leadership skills. Two minutes, fantastic, look, I have two more ideas. It's wonderful in teams when, in teams of people, I mean, when we find ourselves doing, as they generally say, 1 plus one equals three, right? So, from the sum of people, of the team's capabilities, of the skills, we can generate superior value precisely because we work together, we produce collective intelligence, as they say. Here, we use generative artificial intelligence, particularly in this way, with this logic.

Well, all the issues of are we better or are artificial intelligence better? It will steal our jobs, etc. are all, in my opinion, nonsense, in the I mean, in the end, we're different, we're very different things. Um, each one has its own peculiarities. It's up to us to understand what our human peculiarities are and exploit the capabilities of these platforms.

Let's try to co-create with these platforms, not compete with each other. For those who work in communications, therefore, use these systems to communicate, and I think practically all of us do it a little, simply because we ask for help , I don't know, to write emails and therefore to communicate individually. Let's make sure these platforms take our voice, not put their words in our mouths , right? I don't know if you ever see yourself receiving emails, typically messages from anyone or documents that are clearly written by, right? I think this is very common nowadays.

People who maybe until yesterday you knew I wrote in a certain way, poof, suddenly change style, tone of voice, right? It's like a person you know very well, the next day they're dressed, I don't know, in a totally different way. Oops, okay, for goodness sake, but it's a little strange. Here, let's do it. So that the platforms adapt to us and not the other way around.

No, we don't change our outfit just because you have them. Oh, last point, the last leadership characteristic is somewhat the most advanced one, it's the stage of technology we use after having accomplished the first two. The first two are: I do what I already do today a little better, and that's the first thing; we start from there, we go faster, we do things sooner, etc. The second stage is: we do the same things we did yesterday, but in a different way; we change the process a bit . The third is: we invent new things.

Today we have this opportunity, and this is where the competitive landscape comes into play, because the first two stages are being used by billions of people; I think there are now 2-3 billion people using generative AI. Okay? The personal, individual, and corporate competitive landscape is using these tools to do new things that until yesterday didn't make sense because we didn't have the right tools, and now they're innovating. I'll leave you with this, and best of luck to everyone. Thank you.