KI im Mittelstand

Alexander Pokorny: Yeah, and los geht's. Herzlich willkommen, lieber Christian. Ich freue mich riesig, dass du heute bei mir zu Gast im Podcast bist.

Christian Schieber: Hi Alex,

Christian Schieber: So, I'm Christian, I in Nuremberg, I'm computer scientist, but I several companies afterwards, a trip to the startup world and now I'm back in the technology world. I'm somewhere between human and technology, maybe a little more on the technology side and what do I there? I companies, KMUs, also middlemen in the implementation of artificial intelligence. Exactly, so I'll He workshops, also area of AI agents. He interests me very much and I have some exciting projects.

Alexander Pokorny: Very cool. I can confirm that we talked about it You some news about your career as agent. If you want about just the in. I think it's always exciting. Your company is called ByTaboX. We met at an evening event. It was super interesting. You were the host of the event, Future in Beer.

Christian Schieber: You

Christian Schieber: Mm-hmm.

Alexander Pokorny: I can only recommend those who hear and can Buy it now, they will change I the feeling. Let's get The topic is super exciting. We hear so much about KI every day. And KI is most people, I can texts with or write creative texts. I may the Google substitute for some and then creative things, pictures and now

Christian Schieber: Absolute.

Alexander Pokorny: videos. This is the classic world that get from the media. But you care so much more about companies. You said it's about implementations of AI, it's about workshops. Take us with you. What such a collaboration with you? What classic cases do that you can introduce us to? Do you the customer's secret? Maybe one or the other.

Christian Schieber: Mm-hmm.

Alexander Pokorny: for the one the other advantage that you might want to little bit back. That's I would be super happy, take us with you into the world of KI from TurboX.

Christian Schieber: Yes, I like You already the right things. Often you text for text to an email. This is a large language model. You are already in the name. So, rephrasing What we notice is that the mice scratch the surface. This is normal. It's new technology. feel like a new model out You don't exactly.

Alexander Pokorny: Hmm.

Christian Schieber: Which one is good for what? There are countless. And then the companies time to with it. That's it with the simple use cases. You quickly. Write me a new email, maybe a professional larenton, make me some ideas for LinkedIn posts or generate a nice picture. But of course I can it much more. So I don't with it much on this level. And I'm not at every new because 80-90 % are for the German market, or German industry era, we're talking about business, they're useless because of data protection, because the companies aren't in Europe, and then it doesn't But what do we We into the companies, we explain, I work with our friend Julian from One Stop Transformation,

Alexander Pokorny: Hmm.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Mm-hmm. Mm-hmm.

Alexander Pokorny: Genau.

Christian Schieber: We do AI competence workshops together. That means we take a day and really what we with explain bit more, also how good prompt works, not just such a simple prompt. How can you more out of Chatchi BT or other models, but also at ethics, security, data protection, AI in practice. How can I maybe workflows?

Alexander Pokorny: Very cool.

Alexander Pokorny: Mm-hmm. Mm-hmm.

Christian Schieber: or so-called AI agents. Agents are practically like employees, you could say, who for me. And that's where a little more into practice. So that's how we start, I would say.

Alexander Pokorny: Second. Perfect. If you Julian to Julian Knorr, we recorded a episode with him. It was about digital mindset, his absolute non-plus-ultra-thema. He's an expert and he supports you. Very cool. It's a combination you can look forward to. It's super symbiosis. The workshop is the first step in. We want to get it. We may have less experience.

Christian Schieber: top.

Christian Schieber: loot.

Alexander Pokorny: found it very interesting that you said you watch the market but you everything because it's not relevant. And that's just as difficult How should I through all this news, process it, try it check it You can only experts like you or Julian because you very quickly. and also the right things. So that's another plus point why you should definitely work with BiteHabooX, why that absolutely sense. Cool.

Christian Schieber: Abs absolut also genau das also wir geben auch ganz viele Informationen natürlich auch auch kostenlos julian hatten podcast zum beispiel k.i. und tonic kann ich sehr empfehlen auch sehr aus der praxis und wir posten auch viel auf linkedin genau über die dinge die aber eher praxis relevant sind

Alexander Pokorny: Okay, so a workshop with you. What's next? Or what's the result of a workshop?

Christian Schieber: It depends on where the company is. If the company at the beginning, we have participants who have downloaded the and haven't a prompt yet. There is really clarification. Then there are a companies that already it in use. We go one step further and think in the classic workshop style, what are challenges I have in my daily work. You have to say that course now ...

Alexander Pokorny: Mm-hmm. Mm-hmm.

Alexander Pokorny: Mm-hmm.

Christian Schieber: office jobs work much easier than a production. You need more experts for But what are my for my office job? They are and actually really boring. In 95 % of cases people they have too much to do and they already what to do with their time. So it's not about replacing people.

Alexander Pokorny: Yeah.

Alexander Pokorny: Mm-hmm. Mm-hmm.

Alexander Pokorny: Mm-hmm.

Christian Schieber: but everyone. We just to colleagues. They have too much to do and would be happy about more free space that they might to customers for creativity. Other projects that have been the to-do list time somewhere down. That means we look at exactly these tasks, how much time it And then we acquired AI skills beforehand, learned That means we can brainstorm quite classic and think

Alexander Pokorny: Hmm.

Christian Schieber: How can we these problems? Automate, Partly, because automating is often super difficult, even if many promise on LinkedIn. But mostly, I have someone sitting somewhere who oversees But I 60 minutes but maybe I just need 30 minutes We've 30 minutes. If this repeats weekly, will be whole lot of...

Alexander Pokorny: Absolutely.

Christian Schieber: Then we'll about what might be use and has less effort. These are the AI pilots we start quick use with less effort. During the journey we a lot again. With less effort we have such topics as data protection and security under control, as if it a huge AI project. So we'll a few days.

Alexander Pokorny: Mm-hmm. Mm-hmm. Mm-hmm. Mm-hmm. Mm-hmm.

Alexander Pokorny: Mm-hmm.

Christian Schieber: very very good result.

Alexander Pokorny: That's exciting. That's what you want as a company. If you hire you don't just talk about the matter but you can also get into the implementation Because that's what technology promises us. It goes fast. Cool.

Christian Schieber: Absolutely. exactly.

Alexander Pokorny: In an example case, have something in your mind where say, that could be something where most people can a picture of or create it in their so that we can little better what you're

Christian Schieber: Yeah.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Mm-hmm.

Christian Schieber: the model I'm usually a very long prompt, where everything is in and then I can in so-called user prompts. So your brain is practically a big system prompt, and if I to you, hey Alexander, let's go to a and I want pizza, then that's my user prompt. Then you that with what you you and say, hey, I have a good pizzeria in Nuremberg. And that's how we it too. We a system prompt.

Alexander Pokorny: Yeah.

Alexander Pokorny: Yeah.

Alexander Pokorny: Okay. Mm-hmm. Okay.

Christian Schieber: and say, hey, I'm a German company, I'm only in AI tools that work in Germany, are practical relevant, are data protection-conform, whatever. And then can talk to this AI agent or get proactively by email, for example.

Alexander Pokorny: Mm-hmm. Mm-hmm. Mm-hmm. Mm-hmm.

Christian Schieber: and you only your relevant information. So that would be, example, how you could use to only relevant information for yourself.

Alexander Pokorny: Mm. I it very exciting. I about sales just to bit of market screening, so that it bit faster. If you get short update what this week, and then we to the sales meeting, maybe that would great. Cool.

Christian Schieber: Yes, topic of marketing sales, of are not classics. You can a lot of What I exciting, a small project, we a fair preparation, example, accelerated because nobody likes fair preparation. That's something you have to 50 people who at the stand and as time after the fair, after you tired. But you pick up the data of the person at the stand

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Hmm?

Alexander Pokorny: Mm-hmm. Yeah.

Christian Schieber: That means you get an understanding of data processing. And then can, if the person has my message, I them some personal information, simply language or text, again in the information field. And you can say, man, he's interested in our product, he wants some more information about the budget and data protection is very important. And that moment I press send and then he can, or can in this case, AI, my AI agent, this information

Alexander Pokorny: Mm-hmm. Mm-hmm.

Christian Schieber: and can build the email from the individual information I given out. Then the link to the data protection DSGVO guideline to my company. Then maybe the directly from the budget. That means we have a standardized email, but with the individual part. We can course directly, but here again human in the loop. means

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Mm-hmm. Mm-hmm.

Alexander Pokorny: Mm-hmm.

Christian Schieber: Maybe it's enough if the design folder, that means I'm in the train home, just look at these 50 generated emails and them is and the measurement preparation is completed when the train the station.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Great! mean, fairs are getting hotter and The closer you the fair, the more people Typically, people after the fair, because it was exhausting. Or, you because they go back And many companies, it's still really intense pain. It's cool. That's a super exciting case. I think you offer great solution if it works that

Christian Schieber: Yeah. Yeah.

Christian Schieber: and that's not lot of effort. So if you can the hosting complications that you might have, you a data protection. If you don't leave out, it's not lot effort. Yes, I can the in marketing. Then we'll a simple example. Everyone it themselves with their own GPT. So the GPT is like an application in chatGPT. In principle, this is a big system prompt, which I once. And there could, for example, me personas from my customers.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Mm-hmm. Yeah. Okay.

Christian Schieber: and challenge my LinkedIn texts or my marketing texts that I in my brochure or whatever you put there.

Alexander Pokorny: yeah,

Christian Schieber: So that's something simple that anyone can without us. That means I'm just ChatGPT, please a LinkedIn post, but I'll this LinkedIn post and maybe to three or four different people who are my target group. And they should feedback And I feedback again in the next prompt and it. I can the same thing again until I say, now it fit. So if you've it three times, then it's already very, very good.

Alexander Pokorny: Hmm.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Yeah.

Christian Schieber: then you will see that the texts are much better. And that's actually something small, but still the fewest in practice.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: That sounds super easy. Most people have up some personal marketing Or not thought probably they would all lynch They were created. You don't think about They were created and you just in. And then I a digital target group online and say, that would be my LinkedIn contribution. Did I understand What you about it? Or does it for you? Exactly. Cool.

Christian Schieber: This is all.

Christian Schieber: Yeah. Yeah. Yeah.

Christian Schieber: Ja, genau.

Christian Schieber: Exactly. Feedback is simply from perspective of the person. And that can be done very, very easily. And that's also something important, because if we imagine we are the perfect marketing department, then we that out there be And I think with the agency Jung von Matt it was always like that. There was an empty chair and that was practically the customer.

Alexander Pokorny: Hmm?

Alexander Pokorny: Hmm.

Alexander Pokorny: Really? Okay.

Christian Schieber: because everything was against the customer. That was the pre-version of our JetGBT character.

Alexander Pokorny: Okay, I didn't that, but if it was bad, cool. Nice adaptation. That was a very good tip. think everyone can themselves. You can with a normal account. Very cool. Are there other applications where you say, yes, that's cool and it's a bit about writing out?

Christian Schieber: Yeah.

Christian Schieber: Yeah.

Christian Schieber: Yeah, also the classic is actually in the area of business management at the moment. That means most companies want to chat GPT, but are so because the data is in the USA. That means the new ones need data protection conformity GPT somehow. There different possibilities, so the positive news is that works. It is of course bit more than just opening OpenAI. So you can almost every model

Alexander Pokorny: Very exciting.

Alexander Pokorny: Yeah.

Alexander Pokorny: Mm-hmm.

Christian Schieber: Also host yourself, example, on a server. There are also very, good German providers who it well. We that. If I now on my own servers, then I have the big advantage there. I can still maybe own data, because we are now data, we are now data protected. I can my own data. Because what is the disadvantage if I just use JetGBT, if I now company and then I a special case and then I need somehow

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Yeah.

Christian Schieber: documentation, that in my business management, then people sometimes for it for today in the waxes. So everyone who

Alexander Pokorny: Absolutely. Or that they're already because they think it reminds them. I call my colleague and harass him and then there two people who can't work. Exactly.

Alexander Pokorny: Or that, yes. And then again.

Christian Schieber: Yes, exactly. I is case of that is in every company. Everyone who is listening Where was it stored? In which database? And then it was not well ordered. Now there the big advantage. I have my own GPT and I it with data. And what this model can it learn and read than we humans can. That means...

Alexander Pokorny: Absolutely.

Alexander Pokorny: Mm-hmm.

Christian Schieber: I want to look again, so I don't if my machine is broken and then there is a documentation of what I have do. Then I my company's GPT and if these data are connected, then I the right solution. Then it's not just a search, as we know but I can also interact it, as I from JetGPT. That's a bit of a long-term trend right now, because is already lot of information in companies. Of works if I a data-proof, conformal solution.

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: click

Alexander Pokorny: Mm-hmm. Yeah.

Christian Schieber: because I don't to my 5 in the at

Alexander Pokorny: But that's a huge field, because 10 years ago we big data, then suddenly this term disappeared and we started to save to document everything, and now that stuff around and we didn't what power we actually had with And that's big plus of KI, that they these unstructured data or these data amounts. Fortunately, yes, strongly.

Christian Schieber: Yes, exactly. Also exciting in the onboarding Also great case.

Alexander Pokorny: Okay.

Alexander Pokorny: Yes, absolutely. New employees. They into companies as And it's super easy when you say it works like ChatGPT. Then you just have simple functions. I'm just to now. Our language or text input and the communication. Well, we know each from everyday If you're little busy with you know it. Yes, cool.

Christian Schieber: Absolut.

Christian Schieber: Yeah, I mean, there are many other use cases. You can course it HR, also in rig shooting. You have to be a quick disclaimer at this point. How to use it, of course, biometric data. But still, your matching, it's fitting fit for the company, can better. Because the man is not as good in as he always thinks he is. So I combinations there. I can course also in

Alexander Pokorny: Mm-hmm. Yeah.

Alexander Pokorny: Mm-hmm.

Christian Schieber: in finance sector, also with disclaimer sensitive data. But of course, KI can incredibly to make decisions, also company decisions, on in

Alexander Pokorny: Mm-hmm.

Alexander Pokorny: Maybe we BWA and ask what you in it. That's also exciting. Very cool cases. Thanks for the insights. If we want to work you, how do we you? What is the typical contact point where I say, I can to Taboo X, how do you want be contacted by the people?

Christian Schieber: Exactly.

Christian Schieber: So either by email, that works quite easily, or by LinkedIn. These are the typical ways. And then we back chat, see what it's and it has to from both sides, and fit and then it

Alexander Pokorny: Okay.

Alexander Pokorny: Very good. Thank you, dear Christian. Again, if you haven't done yet, click on LinkedIn after and link with Christian. Then you won't any new news about AI in company, use. So really where AI really the benefit is there at TabooX. Great. Thanks again for being I'm forward to many news.

Christian Schieber: You're gonna.

Alexander Pokorny: I'm really that you're in my network and that we can continue recommend TaboX. The middle ground is companies just keep growing and getting stronger again. That would my great passion, my vision for the future. Thank you, dear Christian.

Christian Schieber: Thank you for the beautiful words and the conversation was a of

Alexander Pokorny: Super, ciao ciao.

Christian Schieber: Ciao.