22 min read
Boost Team Productivity with AI: Insights from Quantifly’s Co-founder
Sebastian Schieke
:
May 10, 2024 2:00:00 PM
![Boost Team Productivity with AI: Insights from Quantifly’s Co-founder](https://blog.sebastianschieke.com/hubfs/ep86blog.jpg)
Episode Summary:
Discover how Quantifly uses AI to boost team productivity and solve workplace problems. Co-founder Luka Pregelj shares insights on how AI helps define team roles, eliminate information silos, and create a safe work environment. CEOs looking for practical AI uses in team management will find this episode valuable.
About Luka Pregelj:
-
Innovative Leadership: Co-founded Quantifly and uses AI to improve work and team interactions. Shows his ability to lead and innovate in tech.
-
AI Implementation Expert: Has deep experience using AI to make teams work better together, communicate more clearly, and create a supportive environment. Shows his skill in using AI to solve real business problems.
-
Advocate for Safe Workplaces: Works to make workplaces safer for mental health with AI, showing his dedication to employee well-being and productivity.
Key Takeaways:
-
AI Boosts Team Productivity: AI clarifies team roles, making collaborations more efficient and boosting productivity.
-
Breaks Communication Barriers: AI removes information silos, improving teamwork with open communication and collaboration.
-
Creates a Safe Work Environment: AI helps build a supportive space where teams can innovate and take risks, important for high-performing groups.
-
Learn from a Leader: Luka Pregelj shares how AI improves team management, offering CEOs practical tips for using technology to lead better.
Why It's Important for CEOs Like You
This episode features insights from Luka Pregelj, an expert in business using AI. CEOs face challenges like improving teamwork and communication and creating a healthy, innovative workplace. Luka shares how AI can help solve these problems, improve team dynamics, and support a growth-friendly environment.
The focus is on using AI for better team management. It covers strategies for clearer roles, better teamwork, and a more supportive work atmosphere, showing how AI can help achieve better operations. This talk is for CEOs wanting to use technology to lead more effectively and innovate, which is essential for staying ahead in business changes.
What CEOs Can Start Doing
-
Use AI to make team communication smoother and get rid of delays.
-
Apply AI analytics to define team roles better, boosting everyone's work output.
-
Adopt AI strategies to encourage taking risks and creating a safe, innovative environment.
-
Use AI to improve how teams work together and grow.
-
Turn to AI to lead more empathetically and effectively, and keep up with industry changes.
Conclusion
This episode shows how AI can change team management for the better. It improves communication, sets clear roles, and builds a culture where innovation and support grow. Using AI is a smart move for CEOs looking to lead with understanding and boost productivity and innovation. The steps we suggest are a guide to using AI to improve how teams work and their efficiency.
Chapters
Introduction to Quantifly (00:00 - 02:41)
Leadership Today (02:41 - 07:16)
Understanding How Your Company Works (07:16 - 14:32)
Building Responsibility and Safety (14:32 - 20:28)
From Analysis to Plans (20:28 - 25:11)
Using AI to Analyze Better (25:11 - 31:35)
Smart AI Use (31:35 - 34:30)
Read the full transcript here
-
Transcript
Sebastian Schieke (00:01.06)
Hey Luca, welcome to the show!
Luka (00:04.462)
Well, thank you, Sebastian. It's nice to be here again after more than one year, I think.
Sebastian Schieke (00:08.868)
Yeah, we just discussed, when was the interview? I think it's more than a year ago that we talked about Quantifi and your mission helping enhance leadership in corporates, building a platform. And maybe you can give a recap on what you do, what problems you solve, and maybe give us an update what has changed or basically developed since we spoke back in...
Luka (00:12.718)
Indeed.
Sebastian Schieke (00:38.884)
Yeah, 23.
Luka (00:40.654)
Yeah, sure. So a short summary of what Quantifly is and what Quantifly does is we help managers understand how to improve different patterns of organization, meaning organizational dynamics, passing information, passing delegation of roles and responsibilities and so on, as well as understanding which individuals contribute in what ways, what roles do they play and so on.
and also understanding what is the general climate and culture of the organization. And when we combine all these three dimensions of organization, we get a very accurate model of what is going on. And we can spot specific, let's say, negative patterns of behavior or dynamics and so on, and recommend specific measures to improve them. So this is essentially what Quantiflight does. It goes from...
basic analytics more into prescriptive analytics which we also call diagnostics, hence organizational diagnostics company.
Sebastian Schieke (01:47.812)
And when we look at the typical organization nowadays, I mean, there's so much change happening in the world. We will talk about how you set up your organization. We will talk about AI. So this immense change, this exponential change where every organization is actually faced with what
are the challenges? I mean, organizations, they don't change it fast, you know? They're not built yet to adapt to this fast pace which is happening in the world. So what do you see in terms of leadership dynamics or skills these companies need to develop? And how do you support them with that?
Luka (02:41.998)
This reminds me of, I think it's a quote from Alice in Wonderland. I think the Red Queen says something like, you know, the world moves at ever faster pace. You have to run if just to stay in place. And I think this is super accurate for times right now and the organizations have to adopt to this ever faster environment. And I think that the way to go about it is by first of all focusing on
Sebastian Schieke (02:54.82)
Yeah.
Luka (03:10.542)
psychological safety. You know, we analyzed more than a hundred organizations by now and based on correlation analysis that we've done of different metrics, we see that psychological safety is perhaps the one that most correlates to the other. Obviously correlation does not mean causation, but it's quite clear also from explanatory view that by having an organization in which employees feel safe to...
Sebastian Schieke (03:12.804)
Mm -hmm.
Luka (03:40.398)
safe to take risks, safe to provide constructive criticism to their peers as well as to their subordinates as well as their bosses is what allows the organization to change. So this would be, I think, the first thing, like building on psychological safety. And this is also quite related with another thing that I noticed, which is what is the mindset of organization? Is the mindset or the value of organization to learn?
Sebastian Schieke (03:46.564)
Mm -hmm.
Luka (04:09.678)
or to distribute. Because organizations that are there to learn and create naturally foster a lot of organizational, a lot of psychologically safe environment, and they foster a lot of creativity. Whereas organizations that focus more on exploitation, not necessarily of people, but just they are not focused on innovation quite as much. They are just focused on how do we distribute this amongst each other.
Sebastian Schieke (04:11.428)
Mm -hmm.
Luka (04:39.054)
they tend to focus on responsibilities instead. And that's the huge mindset difference in the two organizations. So I think it begins there with the mindset. And then from there we can go to specific strategies and tactics. But if the mindset is incorrect, then the tactics and strategies will not work either.
Sebastian Schieke (04:59.428)
Yeah, so taking a typical old style matrix organization compared to a startup and seeing, taking this different mindset. And I mean, you realize that, hey, what worked in the past, yeah, with this long decision chains and, I mean, I had a client once and we...
We wanted to improve their sales processes and had some amazing ideas. But it took them about six to eight weeks to get this idea to the board level because they had these board meetings. And then, oh, we didn't manage to talk about this topic now. And we rescheduled it for another.
in another four weeks and these decision times, I mean, you cannot act like this anymore. The market is moving too fast. Your competition is miles ahead.
Luka (06:10.094)
Yeah, I think it's better to implement short iteration cycles where you look at each individual thing as an experiment. Like, you know, we have five different ideas of how we could improve the process. We all admit we don't know which way is the best. We have, you know, different opinions and different ideas, but the best way to do it is to actually test it. So decide on one, implement it, test it. I'm a big proponent of design thinking in that sense. You know, let's first...
Sebastian Schieke (06:14.18)
Exactly. Yeah.
Sebastian Schieke (06:31.524)
Tested.
Luka (06:38.51)
define a problem, then come up with a bunch of potential solutions, and then decide on which one can be tested. Do the minimum viable prototype of whatever it is, and test it as soon as possible. Don't wait. And I think this, in many ways, relinquishes the need for all of these board approvals, because we are not yet committing to the final solution. We are just committing to running an experiment. And instead of being focused on the end result, let's focus on the...
Sebastian Schieke (06:59.46)
Exactly. Yeah.
Luka (07:08.078)
how many experiments have we run? Because if we tested 10 different things, it's a lot more likely that we'll find the one or a combination of a few that actually works. So I think this is a better mindset again.
Sebastian Schieke (07:16.708)
which makes sense.
Oh, I completely agree. So how does Quantify measure or analyze these dynamics in an organization?
Luka (07:27.758)
Well, we get all the data through surveys. So I know nobody likes to solve surveys, but whether you want it or not, it's still the best way to get information.
Sebastian Schieke (07:31.012)
Mm -hmm. We still can't connect to the brains, you know? Maybe in a couple of years we put a sensor.
Luka (07:40.654)
Yeah, I don't know. Neuralink is working on that. I hope that at some point we'll be able to tap into this information. I'm not sure what the European Union and the GDPR crowd will say about that, but I think it would be very useful for that purpose. However, I think that if you simply ask people the right questions, you'll get a lot of answers. And by doing statistical analysis, you can, to a large degree, eliminate subjectivity.
Sebastian Schieke (07:43.908)
Mm.
Sebastian Schieke (07:59.172)
Mm -hmm.
Luka (08:08.174)
You know, if 10 people are pointing in the same direction, there's probably something there. So let's explore this deeper. So the first aspect of how we actually get our data and perform our diagnostics is based on surveys. We do statistical analysis. And the other very important component is so -called sociometric analysis, where we ask people, who do you get information from? Or who do you get tasks from? Or who do you delegate tasks to?
Sebastian Schieke (08:15.14)
Mm -hmm.
Luka (08:36.302)
Who do you work with the most and so on? And with this type of, let's say, relational answers, we build networks, communication networks. And when you think about it, every organization is in effect an organizational, it's a communication network because each individual person acts as a processor. They get information, they process information, they distribute information. That's what they do all day. And so once you can represent it in such a way, some things become so.
Sebastian Schieke (08:51.62)
Yes.
Luka (09:05.038)
obvious that it hurts. Like, you know, we look at the chain of command and the individuals that, let's say the boss said that the subordinate is responsible for something and subordinate says, no, it's the first time I hear it. In this case, the connection will be a dotted line, a gray dotted line. If they agree, it will be a blue line.
Sebastian Schieke (09:28.548)
Mm.
Luka (09:32.078)
And so when you look at the organization as a whole and you see that most of the lines are just gray dotted lines, it tells you everything. Exactly. And oftentimes they come back and say, yeah, but maybe the employees didn't really understand the question. Yeah, that's possible. But maybe they don't understand the hierarchy either. Have you thought about that? Maybe we should work on that and define roles and responsibilities more clearly so that they know where to find information when they need it. And even more importantly,
Sebastian Schieke (09:37.604)
to something wrong.
Luka (10:00.686)
oftentimes in the lower levels of hierarchy, we see people who have like five, six bosses. If six bosses tells you, do this now, it's urgent, how do you prioritize? Especially in organization where psychological safety is not very high and you're afraid to speak up. Or if two bosses tell you to do exactly the opposite thing, what do you do in this case? And a lot of these conflicts that we see as intrapersonal,
Sebastian Schieke (10:07.908)
Mm -hmm.
Luka (10:30.03)
are actually completely systemic. And once you see this in a picture, it's a lot easier to understand. And I find that it's a lot easier for people to accept it as well, because we approach it sort of as a doctor would like, oh, we see this result. It's nothing personal. It is as it is. Now let's talk about potential systemic solutions to these issues that we identified.
Sebastian Schieke (10:54.212)
As a student, long time ago, I of course needed to make some money. And back then, the internet wasn't, I mean, nowadays you can make lots of money online, but back then it wasn't possible. So I was cleaning machines. I was working in a production site and on the weekends, cleaning machines. And I had a team, so I...
I basically went there and I was looking for a student job and the CEO, he just already hired me as a supervisor. Okay, fine, I'll be your supervisor. So I got 15 people cleaning these machines, you know, and I was supervising them. And in the course of a day, on Saturday, I had five managers. I mean, they had five, basically five management team of five people. All these five manager came to this place where I was cleaning the machines. Everyone told me,
a different thing to do. Yeah, it's exactly a great example. I said, listen, guys, you know, this doesn't work like that, you know? I mean, you can't just tell me five different things, you know? And so I basically, I did this a few weeks and then I created.
Luka (12:10.062)
Yeah, and that's, you see, you're a classic example of that because, you know, after that you quit and people start to feel really bad and frustrated in such situations. And the people who are most ambitious will leave first. So, you know, if for no other reason than that, you should really focus on solving this problem. To some degree, it will always be there. You know, we've never found a company that hit any metric at 100%.
Sebastian Schieke (12:31.236)
Exactly.
Luka (12:38.318)
But we found quite a few that hit them at like 93, even up to 96, I think was the highest that we ever measured in anything. So you can get pretty far. And I think it starts with the mindset that I mentioned before. Instead of focusing on responsibility and hunting for witches, like looking who is the responsible person, you should be focusing on where the problem has arisen and how can we prevent it.
Sebastian Schieke (12:45.572)
Mm.
Sebastian Schieke (13:00.676)
Yeah, exactly. Finger pointing.
Luka (13:08.206)
from arising systemically. I find maybe this is an interesting point also on psychological safety, because lately I've been really hearing a lot about it and had quite a few situations where we had to discuss these things. You know, oftentimes bosses or even top management says, yeah, but our people do not take any responsibility. They do not hold any accountability for their actions. OK, sure.
But if you constantly focus on that, then they would rather avoid the blame than take any challenge or take any risk, right? And this is how you kill innovation in your company. And ambitious people, creative people will leave the organization because it's in their ethos, it's in their heart. They need to create. And so if you drive all of them out,
Sebastian Schieke (13:43.748)
Mm.
Sebastian Schieke (14:02.564)
Exactly. Yeah.
Luka (14:04.91)
then sooner or later you'll be just an organization that keeps exploiting the business model that it had, I don't know, 10 years ago, but the market keeps changing and sooner or later you're falling behind, you're dead. And the way to really solve this problem is actually quite simple. You know, give people more responsibility and instead of focusing on who is to blame, let's focus on what happened that this went wrong.
Sebastian Schieke (14:14.82)
You die. Yeah.
Luka (14:32.334)
So focus on the what, not the who. And if you can solve the process problems, then the who is less likely to happen. And the only part where you could or should be looking for a person who is to be held accountable is when the action was malevolent. Because if you imagine this accountability on a spectrum, on one hand, you have pure sabotage. Somebody straight up decided they will destroy your company, and they leaked information or whatever.
Sebastian Schieke (14:59.332)
Yeah, the devil inside.
Luka (15:01.39)
This is your sabbatical. This is the only case where you should be, you know, focusing on individual and finding out what happened here, right? But other than that, human errors are human errors. You're employing humans. So you should be accepting that human errors will occur and how to resolve this? Well, find a better process, implement redundancies to avoid, you know, human error being the critical thing. And then there's the...
Sebastian Schieke (15:14.532)
Exactly.
Luka (15:30.126)
Next, if we move on on this spectrum, after human error, there's what I would call systemic issues. So even the best possible human with the system as it is set up is not likely to be able to complete the task successfully. And a lot of organizations have some processes that are so flawed that even with the best of intentions, people still fail oftentimes.
And that's why I think focusing on the blame is only makes sense when there was a clear case of sabotage. In all other cases, we should be focusing on systemic issues instead. And this is called...
Sebastian Schieke (16:10.788)
which I would argue hardly happens. I would argue that there's hardly a clear case of sabotage in the organization. And now look and ask yourself, how often did you blame someone in your team? How often have you been blamed as an employee or as a team member? So there's real disconnect and I think this is great that you really focus on solving these challenges. So.
We analyze the organization, we do the surveys, we get all the data, we see the dotted lines and the straight lines, whatever. What's next? So how do we turn a dotted line into a straight line?
Luka (16:52.622)
Yes, so in terms of process after we actually got the data and process them, we prepare a report and it's like, you know, 40, 50 pages. Every single metric is, yeah, exactly. Every metric is like explained in depth and also our analysis is there. But just like you said, a lot of people probably don't read the whole document and we understand this, you know, I'm a CEO myself. I don't have, I, I,
Sebastian Schieke (17:02.372)
which no one will read.
Luka (17:21.39)
generally don't have time to read long documents. If I can, I'll go for executive summary. So we really, really focused on executive summary. And it's really simple and short. It's two pages. The first one is the issues we identified and how they connect to each other. And then specific measures you could take to alleviate them. That's it. That's the executive summary. So if you just take these two pages, you already know a lot.
But oftentimes, when you read something, you're like, oh, but how did they get to this? And then we redirect you to specific pages where we explain how we got to this conclusion. Because we don't want to build horoscopes. We want to build actual scientific documents. And so the next step after we deliver the report is to actually also sit down with management and talk to them. And this is a very important part. Because.
Sebastian Schieke (18:01.252)
Excellent. And well, then you had to, yeah.
Luka (18:14.414)
we have to translate this, let's say, observations into the specific context of this organization. And this is where we need their input. Like, have you tried some of these things? What went wrong the time you tried? Or what could you do different this time, and so on? And this can happen on multiple levels. Like, we recently finished working on a large Slovenian company, approximately 3 ,000 employees.
And it's a complex company. It's basically a company that for a lot of legacy reasons has three completely different industries under the same roof. And it's basically 12 separate and relatively autonomous organizations within one organization. So we had to go on basically three levels. First with the board of directors on the very top of the holding company. And then we had to go through...
other levels all the way to the directors and managers of specific production plants and so on and work with each individually to establish a dialogue between these three levels because we found that a lot of issues actually came from you know this misunderstandings or misallocation of information or even lack of alignment of interests within this whole organization so I would say that this part
tends to take quite a bit of time, but it's really, really important to do well. And the end result for us is to build a specific action plan. Our work is not done until the management has a specific action plan. And depending on the issues that we have identified, we sometimes also include employees directly or do focus groups on smaller, let's say, team levels or so on if some teams are deviating a lot from the rest of the group. So it's.
It's a long process, takes about one to three months to complete, but the end result is something really tangible and something that management can execute upon. And based on the companies that already performed the diagnostics multiple times, we can also see the results of this work.
Sebastian Schieke (20:28.58)
Great. So this all sounds very sophisticated. So there's a lot of business logic in your application. And I just asked myself, I mean, we had a discussion before, and you mentioned that you leverage artificial intelligence a lot, and that you set up a process where you have AI agents executing certain tasks. And I mean, I'm also very interested in, and we
We basically also develop a lot of agents for our customers. So what can you give us a bit of a summary? How did you set up your operational and development processes and team? And how do you leverage different AI models and these agents?
Luka (21:14.254)
Sure, I'll try to respond in a way that will be helpful also for any other organization to follow. So the way we did it is in incremental steps. We didn't decide on a solution firsthand. First we decided that we have to improve the process. So our goal was, let's say our KPIs or goals was to...
Sebastian Schieke (21:22.82)
non -technical people.
Sebastian Schieke (21:29.22)
Mm -hmm.
Sebastian Schieke (21:32.74)
Sure.
Luka (21:42.414)
continue reducing the amount of unnecessary human work required for us to deliver the service. So initially, everything was done by humans. We had relatively rudimentary model that was basically just statistical algorithms and so on in the background, and a lot of Excel fights, as you can imagine. So this was the initial. And then for each.
Sebastian Schieke (21:46.852)
Hmm.
Sebastian Schieke (21:54.276)
Yeah.
Sebastian Schieke (22:04.004)
Hmm.
Luka (22:06.83)
individual task we were discussing and asking ourselves what can we do to make this more efficient or is there a way that non -creative human work can be eliminated? And I recall we ran sort of a design thinking workshop internally and our analysts, so these are the people who actually interpret the data, most of them are organizational psychologists, so we talked about their pains.
and we identified their pain being, you know, these reports take too long. Okay. Then we had to define the problem and we got to the point where one of the problems was that they had to manually like read, categorize, and then summarize hundreds and in some cases, thousands of open -ended comments. Can you imagine how dreadful of a work this is? And this can obviously be much better. So even...
Sebastian Schieke (22:36.708)
Hmm.
Sebastian Schieke (23:00.388)
amazing experience.
Luka (23:05.006)
Like at the beginning, what, two and a half, three years ago, so before the GPT revolution, we already started working on some models that could potentially help us alleviate a lot of these tasks. But for us, it couldn't be any better. The GPT models were really a game changer for us. So we, through this design thinking workshop, we came to the conclusion that one of the potential solutions would be to basically train.
train a GPT to do it. We tested it on anonymous data at first because we used a public model just to see if we can get to good enough results and we did. And then in the next step, we actually hired or I would even say partnered with a startup from Silicon Valley that is focused on development of AI models specifically for the kind of companies as ours, so for consulting companies.
Sebastian Schieke (23:43.14)
Sure.
Luka (24:03.694)
and we set up a case with them. So we worked really closely with them. So our tech team was working very closely with their, our data scientists were helping theirs and the other way around. We even restructured some of our databases a little bit to make it easier for AIs to navigate. And the end result was that we built a whole engine that does this work in a matter of.
Yeah, a matter of half a minute or so for thousands of open -ended comments. And then the next step is, of course, creating a GPT that would take all of the existing knowledge that we gather through our data and through all of these reports and basically train it to answer your questions. I don't think we will at any point completely eliminate humans from this process. I think that organizational psychologists bring their...
Sebastian Schieke (24:35.972)
Hmm.
Luka (25:00.11)
creativity, bring their experience and so on to these cases. So it's really great to have them look through this. But it's on the other hand, also really important to only use them where the creativity matters.
Sebastian Schieke (25:11.076)
You don't have to torture them anymore like you did in the past by giving them thousands of comments to analyze.
Luka (25:20.142)
Exactly, exactly. Yeah. Now, if we want to discuss more technical terms, so I'm not a machine learning specialist. I'm not very technically proficient in AI, so I cannot go into too much detail. But basically, what we found is that for navigating these tasks, a single GPT is not, let's say, powerful enough, or it doesn't have the necessary logic.
Because there's a lot of base knowledge that we as humans take for granted. For us, steps one, two, three are very logical. Whereas for this type of engines, they are not. And it's quite similar to managing a person in some way that you shouldn't just assume they understand.
you have to find a way to give them enough space to do their task, but at the same time give them some guidelines in order for them to be able to move forward. And so we were trying and testing different things and we came up with what we call recipes. So we prepared some recipes upfront, like in such cases, try one, two, three, and if this doesn't work, jump to four, three, two, and so on.
And so you basically set it up to perform different operations. And then we created different, let's say, levels of agents. So one agent only takes care of the process, then another agent focuses more on, let's say, mathematical operations, another one more on linguistic operations, and so on. And so they ping pong with each other, and it's almost like creating or recreating basically a team. You know, you have your data scientist, you have your project manager, and so on.
Sebastian Schieke (27:12.964)
Exactly.
Luka (27:15.534)
And so this was the approach that we finally taken. And it was really interesting. I recall the call with the guys from, their name is Psychic, by the way. I think I should plug them here because they were really, really good in this. So as we were working with them, I recall that after we devised this idea of recipes and tested it, it was about a week ago, I think, that basically Google has
Sebastian Schieke (27:31.108)
Yeah, we can put them in the show notes.
Luka (27:45.55)
publish some sort of a white paper with exact, you know, with exact the same idea there and we're like, oh, we are just a bunch of guys and already competing with Google on this. So I think that this is really interesting or valuable to know that this field is changing so quickly that you can be on the forefront with good ideas. It's one of the places where I think human creativity can really play a role. But the important thing here is to try.
Sebastian Schieke (27:49.764)
Yeah.
Sebastian Schieke (28:02.852)
I know.
Luka (28:15.182)
different things and go with incremental steps rather than over committing to a specific idea. We didn't go to psychic first, we tested some things first. So we already knew what can be done and what is gonna be difficult to be able to do. And the other thing, that's why I wanted to plug psychic is it's not obvious to me that every partner would do. Because if they have the mindset or mentality of this is how we do it and that's the only way we do it.
then you cannot work with them really. For us, we needed a very custom -made solution, so we had to work really closely together. For a couple of months, they were basically a part of our team.
Sebastian Schieke (28:56.612)
Nice. But I mean, as you said, this feed is changing so fast. And I recently read an article, I think McKinsey did a study and they said about 50 % of the tasks, general tasks the knowledge workers do can be automated. I would argue maybe it's even 60 or 70, you know? And as you said, building these agents which take certain roles, you can automate.
whole process changing your organization. And then focus on more value creating work for your staff. It's not that you want to replace them, but we want to move these activities which are training and looking at a thousand lines of comments to a machine to free up ourselves, our brains for creativity. And because I mean,
There's so much happening in the world. We all have to process this. We have to adapt, and we have to see, OK, where is the market going? Where do we need to focus on? How can we adapt our business model? We don't have the time anymore to do all the work which we've done in the past. So I think there is a great use case for AI agents for automating activities, repetitive activities, and freeing up human capacity.
and also helping them on a psychological way. Because I mean, it's training when you go. It's like, I mean, in the past, we had production line workers, yeah, they did all the same every day. Now we have people who shuffling data every day, you know, analyzing them. And when we can give this to a machine, I mean, I think that's magic.
Luka (30:48.078)
Yeah, absolutely. Absolutely. Though I would say that we have to be careful about it in a sense that, you know, there are some trends that emerge here and there. And then everybody Johnson bandwagon without really thinking why.
So in our case, we really, we first, we did the whole discovery process where we identified the pain from the pain, we derived a problem and one of the possible solutions was a GPT. It was not the only solution. The alternative was to hire a bunch of students, for example. So we had a lot of different alternatives there and we only decided for this one because it seemed like the most sustainable long -term way, if it works and it was relatively easy to test.
Sebastian Schieke (31:13.668)
Mm -hmm.
Luka (31:35.47)
So I would say that the way to go about implementing AI in even large organization is don't do it top down. Don't decide from now on we are focusing on AI. We will have an AI department that will focus on that. No, don't, don't. That's not the right way to do it in my opinion. I would say that a lot better way is to run discovery workshops with different parts of the organizations and identify their pains and least all of the possible solutions.
And only afterwards, after the problems are clear and some of the solutions basically scream AI, hey, it's really boring. I have to do the same thing all the time. Only then decide to either hire somebody from the outside or build a department inside. But don't jump into onto this bandwagon just because it sounds cool because otherwise, you know, so much money has been wasted on digitalization and
Sebastian Schieke (32:23.012)
Exactly.
Luka (32:32.686)
I see with a lot of organizations where we identify that one of the communication problems or project management problems is supposedly their project management software. And we ask them, but you know, have you tried the different software? Yeah, this is already the fifth we are trying. Well, I guess then maybe it's not the software. Maybe you have a process problem. So let's work on that first. And in one of them, we actually performed a
Sebastian Schieke (32:56.726)
Exactly.
Luka (33:01.55)
de -digitalization. So for a month we killed the software and they had to do basic Kanban on the wall with Post -it notes. And only after this has become adopted, we moved this back to digital software. Yeah, because sometimes it's easier for people to get a hang of or get a grasp of why this is there and how it works. And then Taylor, the...
Sebastian Schieke (33:06.404)
Mm.
Sebastian Schieke (33:15.3)
Implemented Trello.
Luka (33:29.134)
the tool to the process, not the other way around. So this is something that happens way too many times in organizations and a lot of money is thrown away there for implementing things that in the end, piece of employees even more because they know that, you know, there was like 200 ,000 spent on this software that is now a pain in the ass.
Sebastian Schieke (33:50.628)
I've seen many of these projects in the past and I completely agree. Hey, thank you so much. Well, a good update on what we spoke more than a year ago and very interesting to see how you leverage modern technology and how your software is always adapting to the changing world. And...
Luka (33:53.966)
No.
Sebastian Schieke (34:17.412)
Would love to have an hour one and a year's time maybe and see where things have developed to. Is there anything you'd like to share as a closing statement with the audience?
Luka (34:30.03)
I'd just like to thank you very much for inviting me to talk to you again. It's always a pleasant experience. And yes, let's meet up again in one year for some new updates.
Sebastian Schieke (34:40.676)
Thanks for being around, Luca.
Luka (34:42.862)
Thank you.