Introduction
Welcome to another episode of TJ’s Technology Tuesday. Today’s episode was pre-recorded yesterday evening at the hotel, as I’m currently on the road visiting one of my clients. But almost live. And I will definitely check your comments once I finish at the client today. If I get a chance to peek in during a break in the meantime, of course I will do that too.
Yes, but as every Tuesday, an impulse on the topic of TJ’s Technology Tuesday. Last week we looked at the topic of the Hype Cycle – namely the question: What realistic or unrealistic expectations do we have around the topic of AI?
At the start there is always that feeling of: Wow, new technology, fantastic. Then comes the phase of: Oh, none of this actually works. Right now, for example, I am working with a Volksbank where exactly this pattern played out: at first, AI was met with great enthusiasm, then came the first attempts where many concluded it all did not function as expected. And now I am accompanying them in figuring out how to work with what is actually production-ready. Because when even sectors like the heating, plumbing and cooling trade – as shown in this image – are already saying: We are now working with AI, then everyone else had better start doing something too.
So even if Klaus and Ingo are already working with AI, we need to do the same in other industries.
Yes, this then often leads to what I call “tooletries”. Here you can see an overview of the most important AI tools from two years ago. Even back then it was no longer possible to keep a proper overview, and today it sometimes feels a bit like in that cartoon by Ruhe – okay, now we have really lost our way. Also always the question: What actually makes sense? When I look at things like AI-generated 3D figures, I am very much in agreement with Tina Teucher, who says: Folks, when you consider how much water and electricity is consumed for that, it simply does not make sense. What also does not make sense is sending me onto the roof with the best tools available – because there is no socket up there. I always say in this context: Send me onto the roof with the best tools, and you will just have to pray it does not rain. This brings me to my favourite saying, which some of you will certainly know: A fool with a tool is still a fool. In the age of artificial intelligence, I extend this saying with: A fool with artificial intelligence makes the disaster faster.
Why is the topic of artificial intelligence so highly relevant?
And when you look at what the expectations around artificial intelligence typically are – it is a bit like a crystal ball: I have the future, and all around me there is still goose quill and paper, pencil and all the trimmings. In the background the campfire is still burning. And we always think: I just need one AI and then everything will work. It then turns out to be a bit like in this AI-generated image – garbage in, garbage out. Where there are no reasonable data and processes going in, even the best AI cannot save the situation. A good friend of mine, for example, recently had Palantir in the house and they looked at his data and said: Folks, let us talk again in about eighteen months, once you have your data reasonably in order. So why is the topic of artificial intelligence and automation so highly relevant, and why is it important to engage with it? On this point I like to cite a study by the colleagues at AKAD University and Bürokaizen, which showed that approximately 20 % of office time is wasted on searching for information.
Around 30 % is spent on emails, and around 26 % on meetings. So you could say: one quarter is search time, one quarter is emails, one quarter is meetings – and only one quarter remains for productive work. And when I look at manufacturing, they are of course already much further along. But now consider: what would it be like if we ran production the way we run offices? If I were, for example, tightening a wheel with a colleague at the assembly line, and then bing – I run back to the office, then run back to the line, tighten the next wheel, then bing again – I run away again, come back, continue tightening – by the third time my colleague would ask: Have you lost a wheel? In the truest sense of the word. Perhaps it is worth asking: What would it look like if we viewed office processes through the same Kaizen lens? That is the idea behind Büro-Kaizen, developed by Jörg Knoblauch, who sold his father’s company Drillbox and before the sale repeatedly tried to make the business profitable using Kaizen methods. He then sold it while it was still viable and transferred the Kaizen principles to office work.
A second perspective I like to share: Do you have 2,648 letters waiting when you open your letterbox in the evening after coming home? Back in the day, most of us had some kind of wooden or metal letterbox. And if you had 2,648 letters in there – you open it, pull out one letter that is half torn open, put everything back with 2,646 letters still inside. I think if you actually did that, your neighbour would ask what on earth you had been smoking. The advantage of a paper letterbox is that it is a self-limiting system. You have to empty it at some point, and the volume of a normal letterbox is still manageable. It gets harder with a digital inbox of 18,000. You have to clear that out too at some point – but then it becomes considerably more difficult.
In which situations does digitalisation make less sense?
What I also frequently observe is that things which can now be done digitally are being done digitally – just for the sake of it. Like this wonderful example: From now on we will write meeting notes in OneNote. We are digital now. Here I very gladly quote the former CEO of Deutsche Telekom, who said: If you digitalise a cr*p process, you get a cr*p digital process.
Why do I cite him in relation to this process? Because it was never a good idea to capture conversation notes in a simple notebook. It has always been a better idea to look at tasks in some structured form. And with Planner I can do that far more effectively than simply putting automation or AI on top of a cr*p process. An important prerequisite whenever I am called in with the request: Can we not automate our processes with AI? There is always the question of: Can we just have a dental hygiene session? No pain, but nice fresh teeth. I am more the dentist who says: I think we need to take a proper look – there are a few cavities in there and we need to drill. It hurts in the short term, but afterwards you have a brilliant smile again and can bite hard in the competition. So let me summarise: What are the biggest mistakes people make when introducing AI? The biggest mistakes are above all unrealistic expectations, a tool strategy, and bad data and bad processes. It is a bit like what we used to do back at Nixdorf.
For example, when I look at Nixdorf – I joined them in 1988 – Nixdorf used to build their own office buildings, and these buildings always had green strips of lawn around them. The question was: Why does Nixdorf always have green strips around the buildings? Simple – so you cannot hear them throwing money out the window. Then came the follow-up question: Why is there no green strip around headquarters? Simple – because there they are throwing banknotes out. Or the question: What do Windows and a submarine have in common? Simple – as soon as you open the first window, the problems start. Someone who took a different approach was a company in the USA – Mercedes-Benz in America. Car sales in the USA work a bit differently than here. If you walk into your preferred dealership on a Saturday – not one you smoke, but the car kind – the customer typically drives off the lot in the car of their choice that same day. If you tell someone they have to wait nine months for an E-Class, they will ask you what you have been smoking.
But they had a structural problem. What kind of structural problem? A customer sat in a car and said: I love this car, I want it.
But then, as you can see from the image, there were a few exit doors on the way. Which meant that before some customers reached the office to sign the leasing contract, they turned right and walked out. What did they do? They said: at the moment the customer is sitting in the car and says “I love this car”, they present them with: Please sign here. And “Please sign here” means exactly this – and yes, this is deliberately an older example – you can see it is an original first-generation iPad. About as glamorous as signing for a parcel from a courier. But what mattered was the system. The result: a significantly higher closing rate, because at the moment the customer was ready to sign, the salesperson was able to make it as easy as possible. Another company that is very strongly sales-driven is Coca-Cola, and its then-CEO Ulrich Nehammer, whom I had the pleasure of accompanying on part of his digitalisation journey. Number one on the wiki – not only because he is from Vienna, but because he thinks very clearly and states plainly: At Coca-Cola there are two jobs. There is the job of selling Coca-Cola, or helping to sell Coca-Cola.
Speed is one of the most important topics in IT
What he stated very clearly was: one of the most important things in IT is speed. What does that mean? He stood in front of his team at a leadership meeting and said: Folks, today it takes us four weeks from the moment a customer says “I want this refrigerator” to the moment we are able to deliver it. Four weeks in which we make no revenue, four weeks in which the customer makes no revenue, four weeks in which the competition can muscle in. What do we need to do? How do we need to change our processes to get from four weeks down to 24 hours? To achieve this, he said, speed is only possible if we improve collaboration – between logistics, sales and so on. It only works if we achieve connectivity – bringing the systems together – and if we can then scale the whole thing, meaning strategic IT. And that is exactly the point. Recently, for example, I was at a conference where the Marketing Director of MediaMarkt said very clearly: We used to be a box-shifter, and technology helps us today to move from brand positioning to a genuine strategy and brand experience, helps us move from product-centricity to customer-centricity, helps us move from gut-feel to data-orientation, and helps us move from mass communication to personalisation – all the way to hyperpersonalisation.
A small side fun fact, by the way: at MediaMarkt, they developed their own AI voice for advertising spots. And they did not simply purchase some external voice – instead they built an app where every employee could submit a voice sample if they wanted to. From these employee voices an AI voice was then generated. Not just one, but one for each country where MediaMarkt operates. And I genuinely found that idea brilliant. How committed does someone feel when they know the voice used in the ads is partly made up of thousands of their colleagues’ voices, including potentially their own? Truly a remarkable way to bring employee commitment to AI adoption on a completely different level. I really love the whole concept. In other words, what all these companies are doing is this: they view IT strategically. They ask: What financial goals do I want to achieve? And to achieve those financial goals: Which customers do I need to inspire and how? And to inspire those customers: Which processes do I need to have under control? And how do I need to develop my IT and my people to deliver on that?
Conclusion
The whole thing also makes sense in reverse. However, I always say that 90 % should really flow from the top down. Starting from return on equity – the most important financial metric – then down to customers. Which processes do we need to master for that, and which systems and which people do we need to develop? Now there are always new systems – such as AI – where I ask: Interesting, AI – which processes can I optimise with it? How can I use it to inspire customers and ultimately make more money? That is exactly this strategic use of artificial intelligence, the strategic use of digitalisation and IT, that makes sense. Because last but not least – and you know this saying from me – the guiding principle always applies: First switch on the brain, then the technology. And with that I wish you, as your Personal IT Coach for executives, every success from the bottom of my heart.
Yours, Thorsten Jekel.
Key Takeaways
- Unrealistic expectations, poor data quality and a pure tool strategy are the most common mistakes when introducing AI.
- AI cannot rescue bad processes – garbage in, garbage out.
- In the office, around 20 % of working time is spent searching for information, 30 % on emails and 26 % on meetings; only about one quarter is productive work.
- “If you digitalise a cr*p process, you get a cr*p digital process” – clean up processes first, then automate.
- AI should be used strategically: from financial goals through to customers, processes, systems and people.
- Speed is a central IT success factor – Coca-Cola CEO Ulrich Nehammer reduced refrigerator delivery time from four weeks to 24 hours through better collaboration and connectivity.
- Mercedes-Benz USA increased its closing rate by having salespeople use an iPad to capture the customer’s signature right in the car – process optimisation before technology.
- MediaMarkt developed an AI voice from real employee voice samples, significantly boosting employee commitment to AI adoption.
- “A fool with artificial intelligence makes the disaster faster” – technology amplifies both existing strengths and weaknesses.
- The guiding principle is: First switch on the brain, then the technology.
Frequently Asked Questions
What are the biggest mistakes when introducing AI in a company?
According to the presentation, the biggest mistakes when introducing AI are unrealistic expectations, a pure tool strategy without a proper concept, and poor data and poor processes. Anyone who does not address these fundamental issues will not achieve sustainable value – no matter how powerful the AI tool.
What does “garbage in, garbage out” mean in the context of AI?
“Garbage in, garbage out” means in the AI context that even the most powerful AI will not deliver meaningful results if the underlying data and processes are poor. An example from the presentation: Palantir reviewed a client’s data and recommended spending eighteen months improving data quality before AI could be meaningfully deployed.
How much working time is lost to unproductive activities in the office?
According to a study by AKAD University and Bürokaizen, around 20 % of office working time is spent searching for information, approximately 30 % on emails, and around 26 % on meetings. That means only about one quarter of working time remains for genuinely productive tasks.
What does the quote “If you digitalise a cr*p process, you get a cr*p digital process” mean?
This quote, attributed to the former CEO of Deutsche Telekom, makes the point that digitalisation or AI does not improve a bad process – it simply maps it digitally. The recommendation in the presentation is therefore: first clean up and optimise processes, then automate or introduce AI.
What is the Hype Cycle and what role does it play in AI projects?
The Hype Cycle describes the typical pattern with new technologies: first comes great excitement, then disappointment when initial attempts do not work as expected. The presentation cites a Volksbank as an example – initially enthusiastic about AI, then disillusioned, and now learning to work realistically with what is actually production-ready.
How did Mercedes-Benz USA increase its closing rate with a simple digital tool?
Mercedes-Benz USA found that customers often changed their minds on the way from the showroom to the office. By using an iPad directly in the car – with the prompt “Please sign here” – the signature was captured at the moment of purchase intent. This simple process optimisation led to a significantly higher closing rate.
What does Coca-Cola CEO Ulrich Nehammer mean by “speed” as an IT success factor?
Ulrich Nehammer, then-CEO of Coca-Cola, recognised that four weeks of delivery time for a refrigerator was costing revenue and giving competitors room to move. Through improved collaboration between logistics and sales, and better system integration (connectivity), delivery time was reduced to 24 hours. Speed as an IT objective is thus directly linked to revenue growth.
How did MediaMarkt use AI to strengthen employee engagement?
MediaMarkt invited employees to submit their voices via an app and used these recordings to generate an AI voice for advertising spots – one voice per country. This approach generated high employee commitment because every individual felt that they had contributed to the AI solution.
How should AI be used strategically within a company?
According to the presentation, strategic AI use starts with financial goals (e.g. return on equity), then derives which customers need to be inspired, which processes are needed for that, and which systems and people need to be developed. AI is a tool within this chain – not an end in itself.
What does “First switch on the brain, then the technology” mean in practice?
“First switch on the brain, then the technology” is the guiding principle of Thorsten Jekel and means that before any technology or AI deployment, strategy, processes and objectives must first be clear. Tools – including AI – amplify existing strengths and weaknesses, which is why a well-thought-through framework is a prerequisite for success.
Tools & Resources Mentioned
- Artificial Intelligence (AI) – the central topic of this post; the use, mistakes and strategic application of AI in business
- iPad – used in the Mercedes-Benz USA example for digital contract signing directly with the customer
- Microsoft 365 – OneNote and Planner are mentioned as examples of digital meeting notes and task management
- Palantir – cited in the presentation as an example of a company that first assessed a client’s data quality before AI could be deployed
- Bürokaizen (Jörg Knoblauch) – method for transferring Kaizen principles to office processes; basis for the time analysis in the presentation
- AKAD University – study on the distribution of office working time (searching, emails, meetings, productive work)
