Entering the Agentic Age – Leading Across the Great Divide

00:00:00: By twenty thirty, there will be two types of companies.

00:00:03: Those who successfully adopt the AI and those that don't exist anymore.

00:00:07: You have a plane and you are pilot just because your plane gets much better.

00:00:11: then you'll still need to pilot

00:00:13: And this is huge competitive advantage Because other companies do not work like that.

00:00:19: What makes company successful?

00:00:21: I think it's one

00:00:23: aspect

00:00:24: The Boardroom Memo powered by Netconomy.

00:00:33: Welcome to the boardroom memo.

00:00:34: This is episode one, entering the new Chantic Age.

00:00:37: In this episode... ...the CEO of NatConnemy Margin-Bazana and the COO of Natconnemy Martin Wachler unpack what the Chantic AI really is.

00:00:46: Buy a gap between leaders' and leverage's widening And what leaders need to do now?

00:00:50: To be on right side of that divide.

00:00:52: So let's hop into

00:00:53: it.

00:00:53: I've very often been in situation where i thought That was most interesting time for my business life And this is definitely the most interesting time of my business life, because for the first time computers can do something which only humans could do in the past.

00:01:09: I think that's fundamental change.

00:01:11: we don't know yet where it will lead us to and we are just scratching the surface but i agree We can already see so much gain In the end.

00:01:23: It always requires common sense and some logic thinking the most out of that.

00:01:31: I can recall we had a couple points in time where first there was content management, then there were e-commerce and on the channel when everybody expected things will fundamentally change overnight which never happened because people need time to adapt.

00:01:52: And i think it's equally too for AI.

00:01:55: In some cases AI is capable doing much more than we humans, and that also affects us every day in the economy.

00:02:01: Can you imagine?

00:02:03: What is... When you look back what was the point of time or first thing?

00:02:08: did you remember when you noticed yourself that it will change or how I would change something?

00:02:14: I mean..I can recall everybody talking about JetGPT And i had no clue they are talking about.

00:02:22: Then I tried out.

00:02:24: suddenly the machine did something which I could not even explain to myself.

00:02:30: But that was also a point in time where we saw there is huge potential, and it's not so much about what this thing can generate but more like... What you do with that from business perspective?

00:02:49: That kind of rational logical thinking which is biased because every model is biased But in a lot of dimensions can do boring stuff that humans should not do, where it's not more biased than humans are.

00:03:08: And I think that was quite early and there were lots of hype because the majority thought chatting with an AI just generates text will change everything which... There is no reason to believe this happens but from then on I think you can already... And that's also, the strength within economy.

00:03:31: We are doing this now for such a long time and we have always been at the edge of technology That we can recognize a couple of patterns that aren't necessary to make things really meaningful or useful.

00:03:44: So from my perspective there is There is, of course AI that you have embedded into products.

00:03:54: Then there's this personal productivity topic which is now emerging when looking to clothe co-work and gems in Gemini and custom GPDs where you adapt the thing to help do your stuff better or ever more connected.

00:04:12: but I think the biggest potential going forward is if processes autonomously run and just interact with humans where it really matters, so that you can focus the human labor to creating real value.

00:04:28: When I look into customers' organizations nowadays there is a lot of labor fenced in two processes which should not have been done by humans but they lack capacity on other levels Contribute to making customers happy doing things that have evolved the strategy generate new revenue and so on.

00:04:53: For me, when this originally JGBD was launched I was very skeptical because i Have been following what is going in AI or back then it was called people call it machine learning for a long time.

00:05:08: And I was quite skeptical if this hype originated, really was justified.

00:05:15: And originally I used it as everyone you know.

00:05:17: for example i always helped my daughter learn math exams and i use ggpd to check the results.

00:05:25: and you know fifty percent of the timing is correct, fifty percent not...and then sometimes you gave your answers that you thought wow!

00:05:32: That's really impressive..I switched from very skeptical about everything will be completely different two or three times back and forth.

00:05:44: But now for me, last April, so the April hundred eighty five I think is when our first used a genetic coding tools Claude code in Gemini CLI.

00:05:54: then And that from he was really there eye opener how The models and the technology is used how it's embedded in a workflow.

00:06:03: It's really extremely important and can change A lot on what effects are.

00:06:09: And so that for me was really the eye-opener, okay this is very relevant and I have to invest a lot of personal time to understand what's going on.

00:06:19: Also adapt my mindset in workflow into these new possibilities.

00:06:24: as far as hype is concerned it's important not be too skeptical but also not to be too hyped.

00:06:32: Wow, tomorrow everything will be different.

00:06:34: No humans will need it anymore for

00:06:35: anything.".

00:06:37: It's a bit also like the internet bubble in the two thousands.

00:06:42: there was a lot of hype and valuations were overblown at every thing.

00:06:46: but when that stock market recovered you could see there is alot value there and lots things obviously worked.

00:06:56: this would be same on another scale so sure Things that pop bubble whatever you want to call it, but there's some real value There.

00:07:06: You can see and some things that will stay and really completely change how business works.

00:07:11: this my This is my Belief

00:07:15: I think.

00:07:16: what do say?

00:07:17: It's super interesting because they are some.

00:07:20: We can already recognize that there are some some principles that That we can trust.

00:07:28: a true one is that it will never go as fast predicted by those who say that's totally disruptive simply because companies, people need time to adapt.

00:07:39: And the other thing is a lot of impact in adoption are more in the mind of the people affected and not so much technology.

00:07:51: So I can recall conversations with senior engineers here at Economy which explained me their code was better than what the coding agent can create, which is true.

00:08:03: Just a coding agent corrects its mistakes hundred times faster than any human could do that.

00:08:11: and also those models are massively evolving.

00:08:15: but it's always necessary.

00:08:17: you need engineering skills.

00:08:19: I think founder of OpenClaw explained quite well where he said found out doesn't have to write a lot code the engineering skills that are necessary to solve a complex problem.

00:08:34: That also has an effect on business, they're equally same and what I believe is super important for us as our customers.

00:08:43: we build and extend their ability to understand vision of the business model challenges then try support it with smart solutions which is driven by engineering principles, but really gets to the point and therefore it's not necessary that you write all the code on your own.

00:09:04: But you need to understand technology so that you can pull together

00:09:08: stuff."

00:09:09: And I think this

00:09:11: is exactly where you have to understand.

00:09:15: as a programmer or also maybe business person using AI tools... You have to be really skilled in how Interact with the system because that makes a huge difference.

00:09:31: I think you're good.

00:09:31: analogy for that is Let's say You have a plane and you are the pilot.

00:09:36: just because the plane gets a lot better, you will still need a pilot.

00:09:40: We are not in order.

00:09:41: i think industry Is Not Trying to get rid of The of the pilot but actually make it better plane which maybe faster Which we can transport more cargo or some easy roots maybe fly on its own, but you will still need someone to pilot the plane or maybe orchestrate various different planes and so on.

00:10:06: So these skills and technical know-how are still needed... ...and even more understanding how all of those things interact.

00:10:16: then systems work.

00:10:17: this is even more important.

00:10:18: I think It was in the past just how the plane flies and, in our case, how the software has been typed.

00:10:28: This is maybe not as important anymore into future.

00:10:31: until now

00:10:32: I think we can see that with a lot of customer scenarios where we started building agents And their going-in position Was often they already had agents built on a POC Where it wasn't.

00:10:48: consciously They find what is the scope and purpose.

00:10:53: It was more a tech demo than really something which, uh... Which is beneficial for their business.

00:10:59: And then you just fall into hole after technically it works But nobody's using because its doing something that doesn't have much of value.

00:11:10: That also where I see Our organization really growing in understanding the customers need and then come up with a technical solution that's really contributes value.

00:11:19: so we, In future it will be even more important to deeply understand technology.

00:11:24: And what is possible where are the limits?

00:11:27: but also understands business aspect and comes out of best solutions when you create value.

00:11:35: What very often comes to my mind was the statement that Google's CTO saved us in an audience when we were there, in Sunnyvale four months ago.

00:11:46: Which I felt a bit harsh

00:11:48: but

00:11:49: obviously... There is something in it which says by twenty thirty there will be two types of companies those who successfully adopt AI and those don't exist anymore.

00:12:00: And I think that is addressing a super important aspect.

00:12:04: The productivity gain, if you manage to map what's possible with technology especially in the agenetic field... ...with your processes and business model as company will have huge impact on competitive advantage of the market.. ..and it not doubling its tenfolding.

00:12:24: But this has always been case.

00:12:26: somehow You need be competitive and to survive as a company, you have to optimize your processes.

00:12:34: You have to understand the process.

00:12:36: is that you know how things work.

00:12:41: this exactly same now just if say not maybe it's not factor of two but effect of ten or hundred and uh...the difference now.

00:12:51: instead few do not understand your processes in due to no understanding our business actually works and how you don't understand your customers really deeply.

00:13:00: Others will, they'll be not two times but hundred times better.

00:13:05: this is the difference now.

00:13:06: I think that actual underlying mechanics way of business works in these areas.

00:13:14: maybe it's not changing just effects are multiplying a lot because AI such a multiplying factor.

00:13:24: let us look at little bit into The overall mindset that we have towards AI, when you... When we are sometimes in the US and Silicon Valley.

00:13:33: We see that the mindset there is how can be solve problems with their technology?

00:13:38: What?

00:13:38: what can be to benefit?

00:13:40: How Can Be Create New Stuff?

00:13:42: And In Europe It Is Often That We See Even Within Our Company Everybody Starts To Be A Lawyer.

00:13:48: What Could Be The Reason Why I Can't Do This?

00:13:51: Or That can become a huge threat which we need to overcome because when you have agents doing stuff for you within the company, let them do things that only humans did.

00:14:09: Which also made mistakes and where in the end as well is a company fully liable of what they do?

00:14:16: But this mindset shift is something which I think it's super important for us to address and also help educate our customers.

00:14:24: So get things done, allow stuff to happen... ...and not find reasons why you deferred?

00:14:31: How do we see that?

00:14:32: For

00:14:32: me there isn't one or the other approach but you have to find a good balance.

00:14:40: It doesn't just do anything and let AI make all the decisions and do everything without checks, give access to your data.

00:14:50: It doesn't matter which data or machine.

00:14:52: that's not a right solution.

00:14:53: on the other hand locking everything down is not daring because you are afraid someone might say it's also not in the right way.

00:15:03: so I think as company we have to define boundaries really be careful about You give to which system, not only AI but also in every other technology.

00:15:18: Who can see what?

00:15:20: So the basics you have to get right.

00:15:23: and then very important is to have a mindset of curiosity in people end up experimentation because top-down mandated this how we use AI.

00:15:35: I think that will not work... People have to find out themselves Let's say a safe playground to be able to do that without having to think themselves.

00:15:46: Oh, what could happen?

00:15:47: Yeah

00:15:48: I think there is also huge opportunity for you because we understand That some kind of rules are necessary and if we balance it out on the long run can be much more beneficial than just let everything happen And then later Manage damage was foreseeable.

00:16:10: You mentioned data, let's talk a little bit about data.

00:16:13: I think when we look into our customer structure everybody has the data lake but still the data in there is not holistic it isn't always accurate.

00:16:27: semantics are missing.

00:16:29: so what your view on the importance of data going forward?

00:16:31: It's same as before with processes.

00:16:34: you have to understand which data and how different levels of importance.

00:16:44: and the more important, the more basic type of data is for you as a business.

00:16:52: For example financial data.

00:16:54: that has to be correct and it has to right.

00:16:57: You cannot send an invoice through your customer or give some taxes which are about rights.

00:17:07: You have to have that really under control and manage properly.

00:17:11: On the other hand, there may be some operational data... ...some usage data analytics where it might be okay if you just see a trend.

00:17:20: And do you also have to treat that differently?

00:17:23: A lot more experimentation can go on there.. ..you know ,have human interpretation.

00:17:30: so I think this having different levels of granularity or different levels how the company It manages the data from very strict to open.

00:17:42: That's important and people referring back before have to understand which is which, you need access rules.

00:17:51: that has always been like this but gets more important with having AI models giving access to it.

00:17:59: The same if we had some agents you get, give access to the data.

00:18:05: You might first want to try only with this let's say not so business critical things where everything can go wrong and make sure that maybe you simulate before And then have a basic understanding of what are there?

00:18:19: What is the framework in boundaries or what could happen Before really giving access to everything?

00:18:25: Technology also helps alot I think because In the past, always they attempt to have this centralized data lake where you put all of your data in one place which creates a massive latency problem and huge complexity when it comes to integration.

00:18:43: What we also see on our market with lots of customers is that they start building some kind federated data environment zero copy integration, so you don't have to replicate the data.

00:18:58: You just connect it and analyze it holistically which is a super important basis for an agentic layer.

00:19:10: when looking into our ecosystem partners especially SAP and Google that's I think smart move.

00:19:20: they are making ecosystem in one place, and then you just connect the data lakes.

00:19:32: The lack of a better terminology because Data Lake is also somehow outdated.

00:19:37: I have no better terminology but i think it's bit more And You can Have a holistic real-time view on your entire landscape which Is important.

00:19:48: when When I look into what makes companies successful I think there is one aspect which is very underestimated, but where a genetic AI can have huge impact.

00:20:05: Which is the organizational latency that you have.

00:20:08: so sometimes and we see it even within economy when we make decision takes weeks to be implemented just because humans are interacting with humans.

00:20:20: those humans then are interacting other humans.

00:20:23: You can write the rule there, but it leads time to adapt.

00:20:26: But when you have an overarching process which is supported by an agent that agent in a next second can implement changes and the organization is then automatically using it this new way And that's huge competitive advantage because other companies don't work like that adapting to situations where change takes them weeks or months

00:20:51: I think this is what i meant before, can give an example.

00:20:54: Let's say you want to find out why a feature Is taking longer than you expected from the data that you have in The past.

00:21:09: usually You Have some as you said Some Data Lake or some operational data.

00:21:15: Someone builds A dashboard?

00:21:17: And you have To Give it Maybe to some separate department to do all of these analysis.

00:21:27: But if you rethink what is possible now with Gemini or any other model, cloud code for example... I did that just myself!

00:21:41: You don't even need to understand how the data is structured and you don't really have to understand.

00:21:49: Look, please analyze the database.

00:21:53: This is a feature that they are building.

00:21:55: I want to understand why it takes longer than initially expected and really can help you find out.

00:22:03: maybe not exactly but a trend But i would never do that for example when i say okay Why?

00:22:11: As i said before with some really core financial data.

00:22:14: so Some things have to stay really controlled But some things you should have a mindset of, and lot more as possible now without all the effort.

00:22:26: And maybe organizational boundaries and so on.

00:22:31: That brings me to another topic where we had that discussion yesterday A little bit.

00:22:37: Increasing productivity for company requires probably two complementary strategies.

00:22:44: One is define a technology stack, and also it's more than the technology stack foundation where AI and process automation can happen which provides that kind of semantics you were addressing.

00:23:04: The toolset with which you could implement such stuff then requires an mindset when people in company understand It's not just some guys in the IT building something, and you have to specify it.

00:23:21: And wait for a long time to get something.

00:23:24: when you specify it You can also within certain boundaries already build yourself at least To level where you have some kind of prototype Where centralized team Can then take it over?

00:23:39: Get into their enterprise level.

00:23:41: But on a department level, you can do a lot of stuff to increase productivity.

00:23:46: May it be building some kind of analytics or insights or apps like you just explained?

00:23:52: Or even agentic processes that are relevant within the department... And then we are back at the cultural thing.

00:24:01: where the organization needs to understand, you need to give them the tools but also a clear goal and expectation.

00:24:09: That's what we did in economy when said from twenty-twenty six on performance management for every employee which start measuring AI driven impact.

00:24:18: We don't know yet how exactly do that But I can ask couple of questions.

00:24:23: What did you do with the tooling that we provide to increase productivity?

00:24:27: I think this organizational and mindset shift is even more important than technology.

00:24:34: This

00:24:35: also leads me in a net economy, but of course everywhere else people are thinking oh when will AI at some point take my job?

00:24:45: And i think it's completely wrong way for us be.

00:24:51: important is that you learn how to use the tools.

00:24:58: It's for sure more likely someone who can use tools better will take your job and not go away, so it's really important in our economy or also other companies now and it might not be the way that.

00:25:22: It works in two years, but you have to follow it or if they keep up.

00:25:27: otherwise once this new model really emerges there's some kind of agreement point where everything converges.

00:25:39: how do use that?

00:25:40: You have to be ready.

00:25:42: I had to be skilled Otherwise... This is why would you be left behind?

00:25:49: I think there's also a lot of positive aspect in what you just said, because it basically relieves people from doing stuff that humans should not do.

00:26:00: Because the machine can do for you.

00:26:03: if you think back into time of industrial revolution and invention of this steam-machine... ...I guess guys there thought work would go under but we don't know how all will happen as nobody has worked to the biggest growth in history.

00:26:23: Not so good for the planet, by the way but I think from an economic perspective that was absolutely unleashing all the powers and same potential as he also in AI.

00:26:36: it's creativity of humans that defines what should happen?

00:26:44: What is their direction.

00:26:48: And I see by myself, that i can do stuff on my own either making up a mind about something or building some applications at least where you can convey the idea.

00:27:02: Or show a prototype.

00:27:04: That was never possible in the past.

00:27:06: so The way creative humans make impact just massively expanded and there are many good aspects.

00:27:16: also when seeing it in private life When you search for medical advice, your wife is a doctor.

00:27:25: You have direct first-hand access but not everybody has that.

00:27:29: and when I know of people who educate themselves much more it helps them assess the situation correctly And if they are dependent on one person their mood their experience and knowledge to get a single feedback on the super important topic.

00:27:52: There is so much positive upside there, which is interesting when I think about this thing in life that i hate most doing grocery shopping because it's just waste of time but still haven't figured out how supermarket organizes stuff often have to ask for things Going forward a genetic commerce and so on.

00:28:16: What's your view of that?

00:28:18: My my view on genetic commerce if you define it in a way, did use shop not let say yourself in an online shop or over an app But oven agent to do the.

00:28:31: you do the research for products via an agent and son.

00:28:35: my view at the moment is that This will be very relevant For as he said maybe groceries or things where you don't need that much inspiration.

00:28:47: You know, like everyday items... Or maybe I prolong my service contract for whatever which is boring and no one wants to do it.

00:28:59: but these things a hundred percent agentic commerce will be ... And people would do this alot using agents.

00:29:11: But i think what won´t go away?

00:29:13: this inspirational thing where people really like, I don't understand why because i hate it but People really like to either go in a physical store or you know Go online and for hours just look through things And get inspired what they could buy.

00:29:29: So these aspects will not go away.

00:29:31: This may be changed.

00:29:33: You can do quicker Find better fitting products more personalization using AI tools, but I don't think that the main shopping channel in the future will be just say to my agent buy me some cool Nike shoes and it'll do.

00:29:56: Still want to see myself find out what... ...and browse for myself.

00:30:03: this is my view at the moment.

00:30:06: I believe you're right.

00:30:08: we see on market that big ecosystems are trying different things with OpenAI being the first one, with their ACP protocol.

00:30:19: Which I personally believe is very problematic because for a merchant you are just abstracted behind Chatchi PT and don't own that transaction.

00:30:31: You're barely visible.

00:30:35: And it's about product availability and price where I find Kuber's UCB approach much smarter.

00:30:45: And, I can imagine that for research... That has a huge impact.

00:30:51: to make the final decision on my go onsite and do it there?

00:30:56: The amazing thing in all of this is that technology now

00:31:00: available

00:31:02: will be distributed broadly soon Where i can build my own agent.

00:31:11: for me that is costing a lot of time by just taking picture from the fridge and tell it to replenish what I need.

00:31:22: Of course, i would not buy a sofa with that.

00:31:26: That needs more conscious decision And we both know We are not the decision makers with such

00:31:32: business.

00:31:33: Also might be completely wrong.

00:31:36: New generation younger people who grow up do it completely differently, but at least I can never imagine that by a car or sofa in shoes really without seeing them and having some kind of inspiration journey before.

00:32:03: The physical things for sure.

00:32:04: cars want to see how they drive not just have an agent configure it in a way that I like and order without ever seeing before.

00:32:13: And when you think about the other side of this transaction, so our customers what do you think is important?

00:32:23: You talk to a lot of customers...you see how customers think and their company's function.

00:32:33: What are your thoughts on?

00:32:36: Our customers can stay successful and can be winners of this.

00:32:42: a genetic age.

00:32:46: Coming back to what lot people call the great divide, so companies understand how do they adopt AI?

00:32:56: And others who don't.

00:32:59: I think our role in that game is we help out customer make right decisions have the right perspective.

00:33:04: What i see There is so much capital behind this AI bubble, it's difficult for business decision makers to make the right decisions because you get totally biased.

00:33:23: So perhaps we can in a minute also look into how do we look at different ecosystems?

00:33:29: I think that its very important technology and the evolution, things that are possible but also have a very grounded down-to-earth button up view on what's possible.

00:33:45: What really brings value because it is easy to sink millions in AI initiatives with zero value.

00:33:54: I believe we need provide necessary foundation for making right decisions for our customers and help them adopt and learn so that they can control their AI direction by themselves.

00:34:11: I see a lot of very smart, very educated people in our customer organizations but it's difficult for them because usually they get approached by all the different players on the market.

00:34:25: then everybody is selling even better stuff, so it's difficult to figure out what works and what is just a story.

00:34:36: And the lot of that is sometimes just a storey.

00:34:39: .And from that also extracts What Is The Strategically Right Path To Go?

00:34:44: Because Your Investments Need To Be Strategic.

00:34:46: So the investment in twenty-twenty six need to build foundation for Twenty-Twenty Seven ,and you cannot simply replace or change course completely when looking into this, this agentic potential which is there from a business perspective.

00:35:06: There's a lot in the interface to customer and market because on one hand side huge effort marketing sales service whatever And then other hands also huge frustration.

00:35:23: if you for example call hotline of our beloved main airline carrier in our region here, it's really I rather go to the dentist than calling them.

00:35:37: So there is a huge gap and a huge disconnect between what customers expect and what companies deliver on their.

00:35:45: this is definitely one field.

00:35:47: The other thing that you have major ecosystems.

00:35:51: That covers certain topics like SAP for example covering core financial process let us start with and Microsoft covering the Office automation stuff, and for office workplace.

00:36:07: And Google positioning itself on the marketing side.

00:36:13: but you need to bring that together somehow and find the best way how it works.

00:36:18: so perhaps at this point let's look into a little bit our view of those ecosystems.

00:36:25: I'll let you pick your first one

00:36:28: SAP

00:36:29: SAP.

00:36:29: I mean, SAP is from my perspective underestimated because a lot of customers look at SAP as the boring ERP company and unfortunately SAP in their own communication isn't doing much about it.

00:36:46: but SAP to me are only companies that can have the core process end-to-end integrated.

00:36:52: like we see there's lots research on McKinsey.

00:36:58: one-third of the revenues is directly online or it's moving towards that in all major industries.

00:37:07: One third is digitally influenced massively and another third is traditional, so if that's the case then integration between a digital layer that interacts with my clients on the digital field and also generates transactions is super important, so SAP's the only company that can do it end-to-end and I absolutely believe they have not just in their backhand but also customer experience based strongest portfolio way superior to Salesforce.

00:37:38: But there are much better than selling it!

00:37:40: That was one piece within landscape of a company because all those vendors believed that its only them who were here... ...but thats'nt reality, the realty.

00:37:53: You have a mixture of different ecosystem players.

00:37:57: And that's also the challenge when you think about Agentec, it is not just within one ecosystem but across

00:38:03: them.

00:38:03: Then moving on to next what about Google?

00:38:06: Google is an interesting company because they are too strong One technology from my perspective The strongest technology portfolio for all hyperscalars by far.

00:38:21: The other thing is that they are super strong in online marketing and analytics.

00:38:27: So when you're a consumer brand, there's no way to get around Google.

00:38:31: I think it was super healthy for them that OpenAI launched JetGPT because that was the first time in their history where really had to... To think about what they want to be in the future.

00:38:46: And from my perspective In the past twelve to fifteen months They made little mistakes.

00:38:54: And I also think that super complementary to have SAP combined with Google, because you can borrow a lot of technology that you don't have on the SAP side and you can have all those strong business course and combine it with Google.

00:39:11: so that's an interesting play.

00:39:16: a dead end for OpenAI.

00:39:23: That's my personal opinion because whatever you think about OpenAI could do, it is either Google who will be superior or another player which already there and that business model.

00:39:38: I can hardly imagine how they get out of the trap.

00:39:42: And when put yourself in shoes of decision maker one our customers organization who has to decide, you know how the system landscape looks like.

00:39:56: Do I bet on niche players?

00:40:00: Who solve very narrow problems perfectly?

00:40:04: or do i... On other hand of the range choose ecosystem and maybe which cannot solve everything hundred percent as I need.

00:40:15: but What is your view on this, let's say two sides?

00:40:25: We are mostly working for large enterprises which have much more complexity in that aspect.

00:40:34: I think the fundamental question is can i move towards platform or platforms, there might not be one and I can elaborate on that in a minute.

00:40:48: Or... Can i pull in a lot of specialized solutions which are combined smart way?

00:40:56: And they do everything for me but They come with much higher specialization On certain topic.

00:41:04: My belief is since we can build alot stuff our own.

00:41:11: Now, with the help of AI on top of such a platform it's becoming very problematic for a lot of SaaS solutions because their value is substantially shrinking and they bring a lot complexity into game.

00:41:27: We see that even within economy.

00:41:29: if we could find way to reduce half them... ...we would absolutely do that!

00:41:35: And working in it but That's a difficult thing.

00:41:42: Now, when you look into the platforms we haven't talked about Microsoft from my perspective way behind in their IRAs.

00:41:53: I mean You see that also in the investor calls and so on.

00:41:56: they cannot catch up.

00:41:57: A lot of their AI revenue is just proxy open AI revenue.

00:42:06: They are no match in the eye game.

00:42:09: So when you as a company have to make decisions where to bet on, I think there is SAP with all the core systems that are there anyway and the SAP AI portfolio has room to grow.

00:42:30: let's say it like that but that's logical because they simply don't heritage and the capacity to lead that topic, which is fine.

00:42:41: But I think there's this inner security perimeter where you don't like you said before... You want have AI just modify your balance sheet then hallucinate some numbers in the aerosol.

00:42:54: so it needs to be secured And i think than depends on a business model that you have.

00:43:00: In consumer industries Google is super strong player.

00:43:04: I am skeptical that OpenAI has a long-term value because it's too expensive and to less ecosystem.

00:43:18: Consumer industries, you always have your consumers connected?

00:43:23: I think Microsoft has the right to play in certain areas especially office workplace but... It's frustrating to talk to Copilot because it basically is a better help tool, tell me where to click.

00:43:39: That's not what I would expect from an AI.

00:43:42: I personally don't believe that three hundred plus SaaS solutions will be the solution Because there creating so much noise and complexity And cost in end because all of them are seed based.

00:43:55: Then its always better have platform do more on your own Where you can create a lot of output with reasonable effort meanwhile, but the commercials underneath are not seed based because if you have ten tools which are all seed-based You have ten times per seat.

00:44:13: The cost and that's simply becoming too expensive.

00:44:16: What do your view?

00:44:20: I share this view.

00:44:21: what i want to add is maybe the dimension of actually Which AI if it's important which model you use.

00:44:35: I think that this is a very interesting aspect, and i always believed over the last years... ...that its' important to have better-and-better models.

00:44:49: And there are truths of course in how foundation models get better.. ..and they can do alot more things.

00:44:55: But now we see clearly tooling and the systems which you build outside agents, how do you give these agents context?

00:45:09: How this loop works.

00:45:13: The agent does something sends to a model some result gets back it critiques itself changes something in self-improvement loop.

00:45:25: I think that is clear.

00:45:29: It gets more into a direction where it's actually important how those tools work and that they use the models in the right way rather than which model you're using.

00:45:44: So I think this will become more of a commodity for the future, And there'll just be some fixed cost block someone can get but Decided if you are successful or not, which model do use?

00:46:02: Let's talk about agents a bit.

00:46:05: So I think it is also when we talked about agents... ...I often see that this isn't exactly defined.

00:46:12: what that means from my perspective.

00:46:15: there are agents embedded in the software like SAP does with Joule where they're different agents and those agents configurable, predefined.

00:46:28: They serve a certain purpose but they are also open to be approached from outside other agents so the agent can talk with eachother.

00:46:36: and then there is this layer of.

00:46:40: I can create my own agent within local no-code platform where i have visual interface and define an agent that first one is defined scope second has limited flexibility and the heavy lifting needs to go in really this enterprise-grade agent frameworks.

00:47:03: And we see that there is a huge competition going on, Microsoft couple of months ago restarted all their agenda initiatives because they figured out that are not competitive with what it did.

00:47:18: Google was late for party but its great job with their agent development kit.

00:47:22: so Basically, they are also modernizing their whole company based on agent development kit which is really amazing and we can also achieve a lot of results.

00:47:34: But you will see where this leads to because when you look at features like Cloud Co-Works ,which is also a genetic base that becomes the kind of wingman for you in daily business who does stuff while doing other things What do you view on where we stand at

00:47:56: that evolution?

00:47:57: I think it's a very good point.

00:48:02: But, really just in the beginning this is first family of these agentic tools or the first wave of these agents which are under market and obviously already bring a lot value but It's not clear if this is the way that it stays or how does we look in five years.

00:48:27: But what I also meant before, very interesting those tools are those authentic platforms.

00:48:38: Cloud Code, Gemini CLI and Cloud Co-Work actually don't really have A lot of new things, they just use tools which are already there and can combine them in a really good way.

00:48:51: And then fit the results into model, analyze something gives back... ...and you have this improvement loop.

00:49:00: This is an important aspect that you don't need to invent anything new but just use what works in smart ways.

00:49:14: I have to say, that really surprised me and opened my eyes on what is already possible.

00:49:24: But we don't know if the complete workflow of someone working in software development or someone working with finance will stay the same or completely change.

00:49:41: But even more important is to really keep up and use these tools so you can familiarize yourself, be there with your know-how.

00:49:51: It just kills when this next step of this evolution happens.

00:49:56: When we elevate that through the company level I think one thing We already see it's that...we have to build an agentic layer which consists individual agents that follow principles like single purpose and so on, in order to manage the structure of architecture across different ecosystems where you then can start automating really end-to-end processes.

00:50:26: You also need something on top with this user interface for humans to interact with.

00:50:32: maybe I don't see much but something like Gemini Enterprise.

00:50:38: attach all those agents and these orchestration agents know what to do, how to do that.

00:50:45: I think this is something which we can already identify as a good investment where everybody has to follow the path.

00:50:55: so

00:50:56: curious to see what future brings us!