From Sales to Spare Parts - Festo´s AI solutions

This podcast is presented by NXAI, your partner for time series. foundation models and physical AI. Hello everybody! And welcome to a new episode of our Industrial AI Podcast.

My name is Robert Wieber... and my guest today. No it's not the guest It's NOT THE GUEST. I'm in the recording mode with our guests.

Peter Welcome. What do you want me talk about today Robert? Please introduce yourself briefly. Yeah, my name is Peter Sieberg co-producer.

We're together with Robert Weber of the industrial AI podcast. Good morning afternoon evenings to all of you dear listeners. good morning peter. so what do You have peter in a new spot?

Well The big thing Is and it's not news anymore but maybe you And I chat about two three minutes What we think The consequences may be of the United States forbidding non-united states, citizens, the entropic fable top level LLM. What's your first thought? Throw in two three things. I have a recording today with Jakob Tomczak because i want to get the US perspective on the whole topic Because I miss the whole U.

S. Perspective On this Topic In the European discussion. but Im not sure what is Jakob opinion. But yeah That's the word as it is, right?

So that's geopolitics. I did see, I do still read Andrew. He was the guy in charge of the team at Google bringing out Transformer and he's very happy and proud of that. so...

And he is very negative right? Very diplomatically negative about the complete play they have been playing. if you played it this way then thats what will happen. Of course hes big on open source because, you know they didn't allow other companies or builders to use that technology.

And again he's the guy who with our guys here from Germany one of the guys on team brought out transformer. everybody using transformer and these days are technologies as well off course. so we did not like it at all. I think.

for me is this thing about Is there any CEO or any, you know managing director of a smaller company who today now can decide that they're going to continue? To use any not just this. You know any US based AI technology knowing that after having been doing it for A week or months so six months every year suddenly the next minute It Can stop. I think That's The big thing.

Yeah, and it's an invitation to the Chinese models. right because the Chinese are going a totally different approach. A lot of open source open weights models in the market. I see companies using deep-seek in their own ecosystem.

But there is a risk, especially when we talk about small and midsize enterprises right? And I have not an answer yet. so... We don't have the capabilities here to build a model like this.

I don't know. There's an ongoing discussion with Mistral, Le Grand Chat the big fat cat. so there was a meme in the social media feeds about Le Grand Chat and i think they will show a new model present... A new model?

In few weeks but.. But do you have to beat it like exactly now. That's a question or is it perfectly okay to beat him off of you later? you know, isn't the quality of this one.

Of all the Chinese versions and I do not know all of them... Isn't that already good? Ninety-nine percent good enough! And it's really only to one per cent or a five per cent other than ten per cent i don't care.

That is always wanting to be using The very best out there. especially when you talk about industrial use cases, right? So and do see the gap is getting smaller. Right so you launch a new model And then an open source model that's coming in the next four or six seven months at The Gap Is Getting Smaller and Smaller with the performance of the lead time between their leading model under the open-source model.

So let's See. but I think When it comes to Industrial applications You also can go. as I mentioned, I see a lot of industry companies big industrial automotive companies running deep seek models in the back and working with DeepSeek models. Yeah that's potentially another vulnerability there right?

i do very well recall not sure what has been the final decision but you know it was years ago when we were using Hawaii is at telecoms pieces equipment that we were not supposed to. so That's only. you know building on them is only another potential. It's my first reaction today, right?

But when I'm When I visit companies when a visits shop floors where we have a discussion with our guys here in the podcast. I am very surprised and never hear anyone saying oh yes i using mistrial models for this topic. it's interesting Because I recorded an episode was, for example with Jan Seiler from Festo about their whole pipeline buying center at the customer until the spare parts and they are backed by LLM technology. And it's not a mistral topic or something like this.

so there is potential LLM. maybe it's not good enough. The benchmarks are not so good, but we know everybody knows benchmarks of benchmarks and leaderboards a leaderboard. So it depends.

But I don't see a lot of customers in the industry sector using mistry And that's interesting for me. Maybe it has something to do with performance, maybe I have something to deal with a Vendor lock-in topic when it comes the big US model. so its easier to use them and integrate them etc. etc.

Yeah but this could be an interesting question also at AI in Alps why don't you Mistrand models on your devices. Yeah, and then we discuss with Arthur Mench maybe some time or at least exactly going to Paris understand the area that you suggested. I believe mistral is there as well? And may be understand There's a potential that when he got little bit of time from him Exactly We will do it in November With our guys from AI in the UPS and AI on the volcano and AI in Amsterdam.

so we'll visit Paris Exactly. So what else do you have? Alps first, volcano and then Paris. I think it's Federico Martelli.

he seems to be making fun of humanoid startups. He says how to raise around fast. the answer is quite simple found a human robotic startup. so i hear him laughing right Between brackets is the step one that sounds very much like the robotics rules.

There was also afterwards there was a Step zero kind of. have some credibility in to field or build it between brackets. But step one is found a humanoid related startup. number two convince a few researchers from a famous lab.

Number three put AGI and general purpose at the deck. number four invests in good CGI Of One Humanoid Folding Assured or walking, step five raise fifty million plus without ever touching the hardware before. Yeah that's a way to go. now fake it until you make.

Don't take this personal, dear listener who if you and your team just did start one of the humanoids. You know there's always good ones. I don't say bad ones. it is everybody has a right to do whatever they want.

but i think its fair enough. we suggested that before with me when ive go through my track here for hours. There's every day there is tenure, right? And I really hope that in the meantime it must be my best guess around a thousand companies globally.

and then these people who make this split up. where are they? actuators from China but also Germany. I hope that all of you are gonna be right now.

Let me see. There is one additional topic you can access, that's the Claude Founders Playbook for building an AI native startup. so they share how founders are using AI at every stage right? Now of course that only works as long as you're allowed to use.

Claude if suddenly your gonna be cut off. well yeah I need pun intended. now my personal opinion and you may have seen it a couple times copy it or not copied, I'm gonna share more often but here for once. In case you want to have these humanoids work in factories well my humble opinion scaling robots on the factory floors would mean going back to pre-automation.

production was thinking of when did the automation production start? that was T Ford? T Ford was what, one hundred and something. no you could choose any color as long there is black.

Yeah, with the only difference that there's gonna be human. you know it. standing they're rather than humans and Somewhere else. I've said yeah That's that's only going to happen if and then we go back another hundred years.

It happens to be a hundred years as well The original story of the R? u r where the word robot was invented Which is where robots get rid of people? And then sure They can decide what are they gonna do right now unperfectly convinced that instead The industrial robots and the industrial robots, they are going to get an LLM interface or whatever. Let's say AI integrated in what ever kind of form And I'm completely convinced that there gonna have a very good time.

They're gonna continue scaling factories also especially In countries with low robot density. You know like from the United States they wanna do the rebuilding All coming back? Yeah sure But then not gonna do it without. I say without industrial robots, i don't see them.

do it with the humanoid. That's even more difficult. if you need to restart your manufacturing and then restarted with humanoids You're changing two button buttons right? I don't See that happening.

okay As some news from nvidia because he talked about robots They presented halos at automate in chicago essentially applying the concept of automotive industry now robotics. and we already talked about safety in AI and NVIDIA. And I have a guest from NVIDia here to discuss this, they say it's the first full-stack safety system for physical AI... and i will discuss with them because that what Tim Fisher told us.

NVIDA is going directly into this topic now presented as the first concept at The Automate in Chicago. And that's interesting. I'm really looking forward to the episode because i don't understand what they are doing, there is a lot of promotion etc. but the full stack of safety systems sounds very interesting.

so when you talk about Tim from SICK known for safety has he already share at that time, but they are going to be building on top of the stack or is there an opportunity for them? No it's a competitor. It was really an honest question. it's like because you know if there is a company like sick who has been in this market for I don't know, fifty years probably maybe even longer all these typical hidden champions that nobody has ever heard of unless Okay, good.

Well then looking forward to hear from what you're gonna be learning more there. Yeah I'm really looking forward but they were very open to discuss this topic. uh i really liked it and will invite Riccaro Mariani. He's a responsible guy for this topic at NVIDIA And he is a professor in Italy.

Let us see how we get out of the episode because that sounds interesting. I stay with startups for a moment from the guy named Jeff, Jeff Bezos. And he said it's an artificial general engineer for designing, testing simulating whatever the engineer does and building, designing engines, spaying scarf. All of things that you dear listener do in your normal work."

He says,"The idea is to build a set of tools which can actually do engineering". That sounds again very much like what I before mentioned Claude Founders' book right? And he says, an artificial general engineer the dream we've had. He says for decades now who of you listeners then and asking us where Robert is going to be hearing from Prometheus mean?

Where Is what disruption going to take place if it's gonna Be sitting somewhere in the center off whatever The Engineer does design simulation sourcing production all the good things that an engineer does from morning to late afternoon, computer added design PLM. So there's at least companies like Autodesk, Desso, Siemens Altair Rockwell Honeywell and the thousand ten two hundred thousand of The Smaller Ones you are listening. so what is your feeling that do have another thought?

Have you been thinking about What Is Gonna Happen doing everything himself? Is he going to be buying some of the things... He's on shopping too. Yeah, is it really okay?

Yes yes! But he hasn't bought yet or...? I don't know but here has a lot of money in the back billions of euros and billion of euros a big company. and I know some guys from JKU Linz working for Prometheus already.

Oh, really? Yes! When you say JKU it's there...from the university?

JKU University of Linz? yes he hired them For example. Benedict is very smart guy. He worked for NXAI then he worked quite time for Emmy And then went to Promethuis And he was one of the best students at JKU.

So, his hiring people... He's buying companies and you are right. what also Boris told us? engineering simulations software topics PLM Software Topics highly in the main focus I would say Yeah, is it the software that he's interested in?

Or are you really buying data. Are we gonna go back to... We started seven years ago and from day one or two we talked about industrial data. right And its a huge topic.

again this morning I saw everybody starting to understand what happened with Sam including then today's entropic guy, what is his name? Three years ago or really started five four years ago and scraping the internet. And nobody knew at that time they were doing it. now with industrial you can't scrape so the complete industrial automation community, both providers as well as users sit on the biggest mountain of data you can ever imagine.

You know since whatever fifty years that we've been gathering this data and did do it for a passive reason. just said if there was problem somewhere in the field could go back to look at all from beginning. nevertheless an amazing amount am I feeling would almost be like that's what he is buying or must buy. Like all the other companies want to do something, need to buy because you can't really...

You CAN start building and some of them even DO! I think it isn't Noura, they have a gym i believe right? I came from a gym this morning. I didn't see robots sitting there but thats maybe the other way.

That's not a good way, but I think the general way for any other automation company want to do something in industry eye. You can't do it without data right? Exactly and In The moment you have very interesting discussion. when It comes to Germany For example we Have a lot of automotive supplier got bankrupt Etc etc.

And that is Very easy and very cheap To buy them. normally you Buy the chairs and the desks and old machines But there are companies that I'm not interested in all chairs and old machines. They're interested in old servers, right? And so yeah, an old data exactly.

So they really...I'm not telling you the prometos is doing it but their company's outside look exactly which companies got bankrupted then buy the data there by the old servers or PCs etc. very smart. Are you going to be talking to Jacob?

Say hello for me, just one thing. I don't think that's your topic but he says what if AI could write the code But You Could Still Drive? and He Says Ever Since Vibe Coding Became A Thing He Wanted To Have More Control And He Built This Thing Vibe Promptor. Maybe You're Gonna Go Into Details Or Just Wanna Share That.

I Think It'S An Amazing Thing Because i think this approach could mean also for the non-coders to tell the eye what they want, you know in a structured fashion and if it is like today sketching a process. You can talk understands you, or maybe then there is a way. And if this is the way that Jakub found out I thought it was very interesting even more so for non-coders than the coders although Jakub saying when he's coding its a useful thing to do Perfect and we'll ask him. Final thing i have many things but next time is another word loops.

last And when we talked about harness, you recall. We also talk with boroughs about harness... Oh and by the way I had a small story on The Weeknd. again there was a Brewer's Day in Munich.

oh yeah this friend of ours he told us whatever your gonna ask what is Munich to you? people say beer or Octoberfest. that was all about. but now then We call them cold blood, huge horses.

And they always have these cards where the beer is on. and then I was standing with one of those horses some you may be as tourists to Munich seeing him there wonderful. And then they have this harness on their head. I was talking to a friend of mine who does something with steel, and he was explaining me what it was around...

The idea being that the human is showing his horse to go left or right. but at this time i will share my next loop introduced, understand by Boris Charny. He's the creative thought code and he says I don't prompt clothe anymore. my job is to write loops.

now for My feeling it seems like a higher level of telling The agent the LLM agent whatever you interface this to give it a goal at the higher level, right? Rather than a single prompt. So what I typically do...I must say very easy-going guy..

I typically still do the prompts. you know i don't do big things but strongly suggest each of you that are little bit younger then myself To go into next stage and always keep on using AI At as high possible levels. My feeling is its just like I'm a little bit rusty in coding, but there is if then statements right. If value is not bigger than the certain value that you keep on having loop maybe thousand times every second of the temperature stays below twenty five degrees everything's okay and it feels Like That's A Loop But Then At A Higher Level And Thats What They Kill Loops These Days.

Perfect. That's interesting, that's interesting! There is new jobs. well although I don't believe it really i mean certain things...

I don't think there's going to be a loop engineer or whatever but they're certain ways of doing things and everybody's learning from each other. as long we are open-open source kind of thing where you learn from each others so that we see. And if there is, you know coming no jobs out of it like We said the harness engineering That's good But at some point in time It was going to be integrated. and what are you doing?

Well I work in AI. What do You Do? DevOps engineer Loop engineer Yeah, yeah, could be perfect. So for a moment on my side Perfect.

Let's move to the main part and we will prepare each other for the AI in the alps because next week it is AI In The Alps, I'm really looking forward to meet you again Peter. Me as well! Maybe we are lucky that temperatures here are the same which means if we go up into mountains they're a little bit more. Exactly It always like this.

It was a pleasure. Thank You very much. Thanks Robert. Bye bye, have a good day!

Hello everybody and welcome today. I've two guests Jan and Werner. Welcome to the podcast. Jan.

Hi great to be here again. Hey and Werna welcome to the Podcast. Thank you for the invitation. Yes, Jan You have been our guest before.

back then The Gen A hype was just starting. Well come back. Before we start talking about your project at Festo Please introduce yourself briefly Werner. Yeah my name is Werner Reicheltz.

Thirty six years now in the company Festo In the field of automation technology and since five years Now responsible for global sales and business development of our software solutions. So I'm responsible for the complete go-to market approach of Festo. You are responsible for making money, and Jan is responsible to make spending money? Is that right, Jan?

Yes! We're a cost center. I am still heading our department for AI & Control Theory here in Festo Research & Innovation And we have customer facing so developing solutions which actually should help out customers in future. Okay, and we met at an event a few weeks ago.

And I really enjoyed your presentation because you managed to integrate the topic of AI throughout the entire customer journey from the initial inquiry all the way through the replacement part. two spare parts? Could you please explain us what you did in how you did it young? yeah so We have been seeing since awhile already.

that's online sales and also the customer approaches becomes especially for companies like Festo that have many, many parts. Like more than forty thousand parts in the catalog. And then you have two combinations You have all those spare parts. So it gets very complicated For single-sales engineers.

It's extremely complicated to understand All of this. Then additionally we have software solutions like the ones from Perna That get really complicated. We focused on how can we... From the customer idea?

describing that gets to the solution and then even help while the system is running. So from the customer saying, I want a move a box of AA batteries for one meter? And then you ask questions back in her how fast should it be How high? whatever.

You can all ask out of these questions. but you could also include world knowledge From things like another lamb That knows how much one AA battery is and what are the boxes made out of cardamon precisely the right components and combination of components, such that a customer actually gets quickly two solutions at a time to dissolution. The offer is reduced drastically And then afterwards when the customer buys we have predictive maintenance. We also get the right parts at the right times with very smooth experience.

But in the end our goal time to customer solution drastically. So imagine I'm a customer and when have an inquiry too fast, so the first touch point is now in AI or what? Is it sort of first touchpoint? is our website?

sometimes there's also the sales engineer. but if you don't have a dedicated sales engineer The First Touch Point would be on our website And they're your. quickly can go into all virtual assistant and the virtual assistant can help you with Festo inquiries, be it I want to buy a new product or i have problems with the Festo product. Can you help me?

With that and then It's traditional LLM technology are. am I wrong Or is it a rack approach of how You did based on LLMs, of course it's a rack-based solution. We put all our documentation or public documentation that we already have into a rack pipeline and also added some additional knowledge from our application engineers in terms how to sell products for the customers as well as putting this together at the back end of the system. Okay.

And Werner? How reliable is that? In a scale from zero to ten, I would say eight. Eight?

Wow! Because what we did already in the past without LLM was that we provided it with little software tools where they could analyze products and fit them correctly into their application or whatever. so these have been single isolated little software tool for example to dimensioning a cylinder on the right way. And now with LLM, we go at the end in the next steps.

That means LLm is in the background using these tools and so far need and benefit for customers are even higher because they do not have to jump into various software tools So that can just chat with an LL m doing their right thing. But I wanted it also a little bit of outlook Because In future It's not only help our customer finding the right product or huge catalog. We want step deeper and say, okay about AI agents structures we help him even to optimize his complete production facility also going down at the end of digital twin solutions.

To say how can he optimize a complete process not only finding all the right products. but that's an outlook. for We will talk about the outlook i think in twenty minutes. let's keep it.

You're more than welcome. you want to enter podcast now or is there one? I am hiring a podcast completely no exactly. and do what that something yes.

so just wanted to add because we already have this long experience of engineering tools and reduce those in the multi-agent platform, Be sure that the outcome is right. So this is very important, especially for sizing what you want to find a ride motor or You wanna Find your rides pneumatic components just yes out of like thin air. We actually have two calculations. You have to do Like proper engineering and because we use established in four long time established engineering tools here?

We can be very certain That we don't have any hallucinations there but we have like technically grounded answers if it gives good customer. Yeah, that was my question because how do you manage to combine the powerful LLM technology with your engineering tools? With your engineering domain specific knowledge. So as I said we built this architecture...

It's a Knowledge Graph or am i wrong? Well, we have a knowledge graph in the background. So this has been there for quite awhile to do combination of our components. which component fits through each other Component and why does it fit together?

This is a Knowledge Graph that We Have running In The Background For This Combination Combinatory Logic. And Then Now We Basically Have An Orchestrator Agent And With The Chat Front End Where A Customer Can Say What They Want and then the orchestrator agent picks to write engineering tools. They infer from the input, okay now I need electric sizing, ii need chromatic sizing ,Ineed kripper sizing o-ring sizing whatever. we have all of these sizing tools And as Werner said before they were on a page.

you had like fifty different engineering. And now we have like a smart orchestration to pick the right ones, and then the answer comes back. We just displayed that even linked the answer back through the tool itself so you can jump into the tool and change parameters if you want. Before we start recording Jan ,we talked about Mistral and LLM technology in general.

What kind of model do you use there? kind of agnostic, so we can change the model. What we do extensively internally is test the quality of the models because you cannot just change a model and then answers are not correct anymore. So have to have a test set.

let's say proven customer conversations that already has or what from a sales perspective this was requirement in the end proposed as an offer as a test set, so this is not part of the REC process. This isn't part of customer-facing but we use it internally to continuously validate quality for different models such that we can switch quickly and there's no vendor lock in here. So imagine I'm now at the customer, found my cylinder... And now i want to go further!

Is still an AI driven process? Yes or no, I would say. if the customer has then chosen to write product. The next step in the customer journey is crucial.

that means he wants to simulate whether this product Is the right one for his application? So we are now speaking about the face of the final target of a digital twin solution and virtual commissioning. so the full use of AI Will become beneficial for the customer. If they fool Journey for the full machine is digitalized, that means he's simulating whether the machine is running later on correctly and then he's virtually commissioning.

And even programming. there are some topics like coding agents coming into place in so far via a whole customer journey. we speak about AI driven tools but the basis to use them from an automation market point of view, the complete implementation of Digital Twin for our customers in the machine building world. And that's the crucial part which I pointed out also in my presentation.

this is still a jungle and it would have been nice if every component supplier like Festo has all its products completely digitalized. But this is the goal, or am I wrong? This is a huge challenge for the whole automation market because we have thousands and ten thousand of component suppliers all over the world. And hundred-thousands of machine builders in so far will take some time.

now Festo's on good way here but everybody has to do its homework. let say like That's interesting that you are combining the topic digital twin with a sales approach, because looking back five years when Let's use the digital twin to get a better performance during their production. And now you are talking about using the Digital Twin in sales process, that is surprising for me! To add on this correct but sales process as a component supplier mainly...

for the sales process of a machine builder. Because at the end, it's that crucial topic is time to market. That means a machine who was using digital twins Is first off all able To bring his machine faster to its end customer and get them money For then in the past. And thats the crucial competitive advantage.

Then against also our friends The Asian market Who are much more fast than us. Do we underestimate this Topic? A bit I think so. Yes, absolutely because we always look to the OEE, the overall equipment efficiency topic in production but we forgot that machines have to be built first before they can produce Because if they are ready to produce The construction phase and design phase of a machine is already over.

But this is big problem against Chinese competitors that machine builders in Europe and Germany are too slow. In implementing their machines, then bring them to the market. Jan coming back two To The Customer Robert. so now I decided this cylinder.

what is it? Can you describe a little more of process? What's happening Now? if You Decided Would You Want to Buy It And You would need to Generate an Offer and at the Moments Dare the Customer Journey as it Is Implemented Which Lead You to our Fox website where you then can start to order and You have a bit of material, and then you can basically Buy it.

And as Werner said the goal would be To actually take all The knowledge that you have from engineering save It somewhere and include into your project. This is not yet there this vision. so at the moment There's a bit Of breaking point. you Have all of this deciding what you want to put Then buy it get it delivered and you have to mount But setting up and set-up in minutes starts.

So this is another goal that we do on the customer journey, where then also again use AI for example. if it's not a cylinder but an electric axis you need to program it. You need find the right control parameters And there are many little helpers included either in our tools like the Festa Automation Suite or include into component itself so such that they can auto-tune to the rest of environment because as a component supplier we often don't know how our components are used, right?

The customer buys it and then put in some machine somewhere. Sometimes you'll notice very often when not knowing about it but still has work. It's reliable or set up as easily as possible. We need all these little tweaking and auto tuning models.

it helped the customer to find optimal performance for its use case. How do you do that? Can you give me an example about this AutoTune model? Yeah, so one example of auto tuning which we already have today is if you buy a faster motor but you already have another supplier axis and you want to optimize the parameters of that motor such as the axis moves smoothly path planning works and so on.

There are some parameters from the axis that we don't know. friction, We do not know the exact size of it in those many things if you just connect to motor is a mechanical connection but there's no data connection. So we have to identify those parameters while this system is set up. This one example called autotuning.

The idea here is you have a fester motor with external access. You can connect them. Then you make a training run, very simple training one. You identify the important parameters like your friction and your mass...

So in context learning? Exactly! And then you tune the control parameters of your motor such that you have an optimal movement. Okay okay That's interesting.

When I now need this spare part when i'm coming back to the AI or is it a typical service approach? There we have a hybrid approach. So there are two paths that you can do, You can go back to the AI and say I Have a problem setting up. Can you help me?

And then it would again Do direct process look into our documentation and point out To correct things for your to-do. either Help you with this setup. or if you say i don't know this LED blinks What does this mean? This actually means That your cam pullup Does not work something like that.

so It links. But this sounds Not rocket science but its still not solved very well. Or am i wrong Jan? Because everybody's talking about this use case, right?

Yeah. I think we solved it pretty well. so We give out the answer together with sources and you can read where we found that information And then we have documentation Together with our typical Q&A That we already had for typical customer questions Mm-hmm. What else is possible there now?

Let's talk a little bit about your outlook when it comes to spare parts, When It Comes To Predictive Maintenance Still Not Solved From My Point Of View. what else do you think Is Coming To The Festo Components? coming back to this topic digital twin it's mainly the topic of digital product passport which also has a legal background. and at the end because you as a component supplier have to declare hundred percent where does your spare part?

or what are the components? the digital twin sub model, which is called Digital Product Passport. Is at the end a crucial evolutionary step also for us because we are able then to link our customers closer to us Because from our component they get all their information. so that I do not just say okay i take this spare part of another supplier because They're faster in delivery or whatever.

So it's At The End. the Digital Product passport gives Also A Chance To Inform Our Customers Proactively For Example about phase out over for product or technical new features coming with the product and in so far this is a next step in customer binding I would say to using AI and digital twin for that. Jan, come back to the AI topic. prediction not solved yet?

Or am i wrong? well there are very strong prediction models for predictive maintenance. Remaining useful life is still active research, so there are some approaches but none that's established as like the one approach that works for everything and definitely things to do. I think what we have been even before the agentic and gen AI hype pretty good in this predictor health stage right?

To say how looking normal, like this anomaly detection. But actually making good predictions and saying okay it is in three weeks or in like... So make an anomaly detection on the forecast right? Yeah so that's still open.

yeah This was interesting research work we're working on. Okay I want to come back Because we already discussed this topic. What about KPIs? About your customer journey, AI driven customer journey.

can you share some KPIs Jan? Yes sure so one is how often do customers come back? Like if a customer uses the virtual assistant, does it come back to the Virtual Assistant? And there we have return rate of eighty seven percent.

So most of the customers that use the virtual system came back to it. so We assume they found useful. Of course There is typical. you can make thumbs up If this answer Is useful mechanism but very rarely people Use that.

Basically nobody uses That! And then the next step, of course is from the answers that were generated. Did this actually lead to sales? This also looks promising but it does not yet completely dare I would say.

so there we still have to improve. There are people that actually jump to it But very rarely somebody From the solution directly jumps too. I buy this. So typically you use this as a first like.

this is a good idea. try to get in call the sales engineer exactly then you're getting contact with the sales engineering. is okay this? So there is some room for improvement.

Werner, you want to add something? In sales the KPIs of course are very simple. at the end it's turn over but a good KPI to measure whether our journey is working. this for example the turn rate from quotation-to-order.

that means if we complete the process in an early stage and sending then for the product he has selected or wants to buy. Then we can follow up this quotation, also AI driven by way and say okay did you really order based on that quotation? Or not? And so far We Can make your feedback loop.

then back To The Customers. Behavior Jan mentioned the sales engineer in. it's Not all about technical issues But Also About The Organization. Werner I know Sales Engineers In The Automation Business Since Fifteen Years And they are not so keen to sell digital products, software products.

How did you get your people in the different departments on board this AI-driven approach? We have to distinguish between business we're doing on a project oriented basis and business we do on standardized basis. that means our principal approaches is to sell software solutions like hardware solutions which mean complete... encapsulated solution which has a part number and can be sold like a piece of cylinder or piece of valve.

Insofar, this is the one approach to come into a scaling business approach on the other hand side that digitalization market. So our sales engineers and application engineers have to be trained systematically to handle such projects. And that's a challenge in training them regularly, software solutions for data science applications or whatever but this is also ongoing. But do they see the benefits, right?

They see the benefit of course. But at the end coming back to our original business model is we are selling many components and customers. so first off all a software as part or offer in the automation market. Software has nothing else than a piece of tube.

that's the principle approach. otherwise you don't come into quantities. but on the other hand We earn our money and in future also, we will own a money by selling hardware. So selling software is only serving the hardware sales more or less because the majority part of our turnover is done by hardware sales.

Of course. Let's have a quick outlook. What does on your AI agenda Jan? In terms of customer journey The customer journey and then in general.

Okay, both so for a customer journey we are enlarging this multi-agent framework that were adding more engineering tools. We will also look into cross technological engineering tools In the past where like one four dramatic And than one fire electric and you can kind of cross reference them But to calculations I still run separately. and as fast though is looking into offering really seamless automation over all technologies. We need more of this cross-technology engineering tools.

And then also looking into the components, this is for troubleshooting that we said like read out error codes automatically derive tasks that can be done from there and maybe even make automatic maintenance task to put them in. things such as our smart enhance tool so you could automatically schedule your maintenance workers. this is in your error code, you should exchange this and add the spare part. In general we are looking into making our components even better so adding AI features pressure control, we will have infosition detection.

We have system identification and all of this to make again the usage of our components easier And what Werner said in a very beginning To Make it possible that The whole System can optimize itself so That they Can talk to each other and see okay here I Have too much load or i could even move faster and things like that and add that In the Components using Our component knowledge to really embed it in the firmware and make our components even more useful for customers. For that you need standard applications, standard APIs?

That's a crucial point. as I said before The problem is that a machine consists not only of FESTA components. Exactly! That's the dream, yeah?

That would be the dream for every sales engineer at first but in the end several hundred component suppliers forming complete machines and the problems with digital twin approach which are the basis to all this. as Jan pointed out... the problem was that a component from us cannot communicate with a component or somebody else if there are no standards. That means how is a simulation tool working with the other simulation model from another supplier?

In other words, it's not an ecosystem. So putting together of five simulation models will not form the simulation model for entire machine and that still big problem. so they have no same language. as long we are in our eternal world, festival.

that's easy and this we will do. And there we are for sure on a good way but the machine builder has no benefit out of it. if we're doing our homework completely... Exactly!

and another component supplier is not doing so. The workload is again in his shoulders as today already. So you need to enable the machine builders to use AI? Yes, they do already use AI.

But AI is only so clever as data coming from single components... Exactly! from digital twins and sub-models like digital product passports etc. And then AI can handle this data.

but if these are in a different format or not following strictly for example E class parameterization then it's very, very difficult to put just together five or ten different submodels and getting out the machine by mouse clicks. That is not possible. so there are always a high workload on the machine builder side that uses AI because models do not speak with each other in the right language which at this moment is crucial part What is your AI agenda from a sales perspective Werner?

Yeah, we are working with AI also in different areas. Also to train for example all sales engineers. so we have done AI driven avatar solutions to train our sales engineers for different sales situations and partners such things. Then we are using AI also in terms of marketing and preparation for our markets.

Analyzing the whole, let's say user traffic on our homepage bringing all these things together. there We're doing a lot already but here is the same problem. The last meter to close this deal Is always the most important one And still thanks God humans. They want to negotiations with Festo.

Exactly! So if two AI agents are negotiating our products, that would be nice but I think it will never happen. Okay Jan looking back last four or six weeks what fascinated you when it comes to AI? What was something you said?

oh this is interesting i should share. Okay, so four to six weeks. I think of course everybody's looking into how scaffolding is evolving and all the new skills around it. So i think this is fascinating.

but my personal interest is very deeply in two differentiable simulations. See that This has been growing slowly But steadily. And if we look why AI Is so successful That it is today It boils down largely To data that is differentiable. And if we use that logic in other fields like simulation, you can use that success maybe and other things.

Like developing components or developing your solution or control parameters. so this is something my interest lies very big of course. T-Rex was also a nice surprise to have the new release pushing time series forward because this is a big modality that we have in industry and I like that also very much. It's interesting what happened since two years.

we had the last recording. Jan, it was a pleasure to talk to you again and it was really a pressure too at Werner because he is The guy who needs to sell all this stuff You are developing And here's the proof At the end that the customer Is happy. Thanks A lot guys! It Was a Pleasure To Talk to All the Best.

Keep Our Fingers Crossed For Festo & Your Component.