How the Rolodex was digitalized and what you should learn from it

There used to be a time, which I hope many of you have forgotten about, that the Rolodex ruled the desk of every manager around the world. Being in someone’s Rolodex and having important contacts in it, was a sign of accomplishment in the business world.

Since then, this analog way of storing contact information has moved into the museum as digital alternatives have completely taken over. The process of how the Rolodex was digitalized, provides important lessons we can put into practice today when digitalizing our business processes.

What first came to mind…

The first version of a digital version of a Rolodex I used was Outlook 95. It allowed to save contacts, along with the same information that people would put on their contact cards such as their address and telephone number. And of course their e-mail address, although almost no-one had an e-mail address back then.

At the time, I thought it was brilliant. Searching for contacts digitally was much easier than physically.

Looking back, being honest, it really wasn’t an improvement. I had to type over all the information on every card I got, instead of just sticking the card in the Rolodex. That lead to typos, which meant I had to call the company the person worked at to get the correct information.

Also, this process took way more time than simply adding a card to a Rolodex. And while we’re being honest … searching digitally was cool, but back then not really much faster than the Rolodex…

Digitalizing analog processes

The problem with digitalizing a process in a straigh-forward way, like the example of the contacts in Outlook versus the Rolodex, usually leads to less efficient processes. There are inherent differences between analog processes and digital processes. These differences can either be good or bad, depending on how you utilize them.

Analog Digital
Paper is hard to copy Digital data is easy to copy
Easy to secure Hard to secure
Hard to access data Easy to access data
Hard to scale up Easy to scale up
Easy to change process Hard to change process
Hard to ensure compliance Easy to ensure compliance

Let’s take a quick look at some of the aspects and how they can be used in a good way and a bad way. For example, paper is hard to copy. This can be an obvious weakness when we are trying to provide information on a contract to a different location of the company for example. Digital information can easily be copied, for example simply by mailing the document to a person that works at the other location.

But you can also leverage this aspect of a paper process. Things being hard to copy, means they are harder to steal or leak for example. A signed physical document is much harder to falsify than its digital equivalent. Similar arguments can be made for accessibility and security.

Even compliance being hard can be a positive. Paper processes are notoriously hard to comply with simply because there is not a real restriction on executing the process. No matter how specific your paper form is, someone can scribble something extra on the paper or fill out a field completely incorrectly.

For digital processes, these steps can be made in such a way that only the exactly correct information can be entered. In many cases this is a positive, but there are cases where this leads to edge cases not being possible. When the process is very strict, it is easy to comply with, but it loses flexibility.

What today’s digitalized Rolodex looks like

Today’s version of the Rolodex completely leverages the advantages that can be achieved by going digital. However, it also has some weaknesses that come along with going digital. Weighing the pros and cons though, I would say that LinkedIn is a huge improvement over the analog Rolodex. But it works in a completely different way.

LinkedIn leverages the fact that digital information can easily be accessed and copied. Instead of relying on every person to maintain their own copy of someone’s data, LinkedIn simply allows you to link to the data of a person.

This means that when the person’s phone number or address changes, it is effectively copied into your contact list. Much easier than the digital Rolodex example I gave earlier.

Because LinkedIn is digital, it is very easily accessible. This means I can access my digital Rolodex from home, work, while on the road, and via any device that I am using at that particular moment.

Accessibility is one of the most powerful aspects of digital processes. But it becomes even more powerful when combined with the ease of copying data, as the LinkedIn example shows us.

A big downside of digital, is that it can be hacked. People can steal your password, or they can hack into the system (after all, it’s easily accessible), and get your personal data out of it. This is a big downside, admittedly. That is why we are always stressing security so much when considering digitalizing processes. And security is hard for businesses, as witnessed by the many data leaks we are confronted with every week.

Digital is better

All things considered though, I wouldn’t want to go back to my old Rolodex. LinkedIn has many benefits added to my old Rolodex that were simply impossible in the analog age. Perhaps you found this article through LinkedIn because I or someone else shared it. Or you found a new employer, business partner or customer through your 2nd line or 3rd line connections.

These are all advantages that were impossible before the Rolodex went digital in the right way. So when it comes to digital processes, always remember to leverage the correct aspects of digital to make sure your new process blows the old one out of the water!

How to draft a plan of improvement

In our previous blogs, we discussed the importance of determining the happy flow and addressing the bottle necks. Once you’ve asked all necessary questions, you can take concrete steps towards creating your own application. Key to this process is a nicely timed series of ‘what ifs.’

Alternative scenarios, subsequent steps

What if we would leave these process steps out? What if we would add another step? What if we would create an application that sends text messages in addition to emails? Such questions allow you to broaden your horizon and come to an actionable plan of improvement.

However, it is essential to write down all subsequent steps, because one single change in your process will likely cause a so-called ripple effect: many other steps will be affected by it. Therefore, you need to check if this change will not result in any problems later in the process and if it will create a desirable overall outcome.

By way of example, let’s have a look at the above mentioned text message solution. Suppose an employee with a commercial position receives tons of emails each day, which often causes new account requests to end up on the bottom of the pile. This, in turn, leads to unwanted and unnecessary delays.

So, you could try to create an application that first sends out an email, but that also sends text message reminders at set intervals. And if said the employee does not respond to the latter, the application may ultimately send the request to an authorized co-worker.

Now, it is important to ‘tell’ yourself the subsequent story: write down exactly what would happen next, taking the smallest steps into account. That way, you can see if the application you have in mind will work in practice!

Making decisions, drafting a plan

If all the steps lead to a desired result, you can definitively opt for a specific solution. When writing down your plan of improvement, be short and concise: in most cases, one page will suffice. Using this plan, you can create the first version of your application, which you will continuously improve.

The reason why it’s important to make something tangible quickly is that once created, your application will allow you to make improvements by trial and error. If it does not yet yield the hoped-for result, you can easily go back to your series of ‘what-if questions’ and adjust the application where needed!

Keep in mind that changing a lot at once can make your improvement process very slow. You’d want to offer your users new features as soon as possible, so it’s recommended that you divide this process into phases. Keep an eye on our blog to read more about this subject!

The road to success: finding the bottlenecks in your process

“Where can we make the greatest impact?” It’s a question that sets you off in the right direction when automating your processes. Once you have selected a problem or process to start with, you need to get all your stakeholders on board.

After all, you want your co-workers, customers and/or suppliers to quickly see the added value of the application you wish to make. To achieve this, addressing the biggest bottleneck helps you to create support for your digitalization endeavors.

Reveal the bottleneck: get to the root of the problem

Where does it often go wrong in your process? What action costs a lot of time? What is a major point for improvement? Usually, it is necessary to keep asking questions for a while. So go out and do interviews and create a brainstorm session, to reveal a key bottleneck within your process.

Moreover, it is important to align opinions on the subject matter: three people may have three different answers to the same question, as they might experience a process differently.

Once you’ve decided on a major bottleneck, it’s time to focus on the ‘how’ and ‘why:’ How do you carry out the process and why do you do so in this particular way? By asking these types of questions, you discover the root of the problem.

This allows you to look for a solution. Remember that the most logical one isn’t always the best option. So, join forces and brainstorm away to figure out what solution best fits your organization.

An example: on-boarding new employees

Is the above still a bit abstract? Here’s a practical example: the on-boarding process of new employees. Usually, this process starts in the HRM department when a vacancy is posted. Once a new employee has been selected, certain actions need to be completed before he or she can start working for the company: a new laptop and cell phone must be arranged, and the new employee will need their own account to access the company’s data.

Multiple people and multiple departments are involved in the process, and any delays cause the entire chain to stagnate. If, for example, the new account is requested through email, it ends up on a large pile, where it is bound to be forgotten. This can be a major bottleneck.

Now, if you build an application to control and monitor the process, that automatically sends reminders and forwards the process to the next person responsible upon completion of each step, you will instantly see a huge improvement – in terms of both speed and efficiency!

Steps towards determining the happy flow

If you want to jump on the digitalization train, you will quickly find that it’s key to define your own process before turning it into an application. An effective way to do this is by asking the right questions. In this new blog series, we will guide you through this process step by step – starting with the happy flow!

What is a happy flow?

What does your process look like when everything goes the way it should? If you answer this question, you will get grip on the core of your process. We call this the happy flow. This is the most important thing to know about your process because it will keep your initial digitalization project small, completed on time and it will prevent loads of double work. But how to go about it?

#1 Start at the end

If you first describe the end – or, the desired outcome – of a process, it’s much easier to define its starting point as well. The end result could be a quote or a document of approval. The next step is the starting point. The latter can be an employee taking initial action, a customer submitting a request, or a supplier specifying a certain need, for example. Whatever the specifics of your situation may be, it’s important to first determine your end result and starting point, respectively.

#2 Determine the in-between steps

Now, it’s crucial to keep asking questions about the process, so you can fill in the gaps and reveal the in-between steps. By doing so you visualize both the happy flow and the exceptions in your process. The easiest way is to use a whiteboard or a notebook, so it’s easy to connect the dots and visualize your process.

#4 Components of the happy flow

There’s no need to do this chronologically; simply jot down every step or action you believe to be standard and important to the process.

#5 Exceptions

Whenever you feel that you’re describing an exception rather than a rule – for example, a step that’s only relevant if the amount involved exceeds 10.000 euros – you simply write them down on the exceptions list, you will need them in another stage.

#6 Create the entire flow

Once you’ve distilled all the standard important process steps, list them chronologically on a piece of paper. That’s your happy flow: a useful overview that provides insight into your process. This will serve as the framework for the first version of the application you want to create. When this application is tested and approved you can decide if and where you want to add steps from the exceptions list.

Curious about other important questions you should be asking? In other blogs, we’ll tell you how to use this approach to reveal the bottle necks.

Implementing a digital strategy despite the immense skills shortage

Every year I get the CIO Survey by Harvey Nash / KPMG. While the full report isn’t available until June 26th, some of the key findings are. In my opinion, of all the key findings, one in particular stands out. The immense skills shortage.

Skills shortage

The CIO Survey is conducted amongst about 4,000 IT leaders worldwide. Therefore, it gives a very adequate overview of the current state of businesses around the globe.

And while there are many interesting facts every year, the skills shortage is one that has been steadily growing over the past years. This year, 65% of all respondents claimed to have a problem with finding the right IT skills. This is the highest it has been since 2008, which is very alarming indeed.

The result of not being able to find the right skills leads to difficulties implementing a company’s digital strategy. It is therefore no surprise that 78% of all respondents said their digital strategy is only moderately effective, or worse.

Especially taking into consideration the fact that failing to correctly implement a good digital strategy is one of the main reasons companies fail to compete nowadays.

Let go of the faster horse

I believe that companies should actively look for different ways to solve the problem. Many are still trying to find more specialists, where there are none.

While it is tempting to think that we can get everyone to code, the simple fact is that it takes a certain profile to like coding. Just like it isn’t for everyone to be on stage presenting to a large audience, coding and other technical skills simply aren’t for everyone.

Taking this into account, managers should actively search for different kinds of solutions. Henry Ford famously said that people think they want a faster horse, while they really want easy transportation. In IT we should stop looking for that faster horse and instead find other ways to reach our goal.

One way is to look for solutions that actually allow a different type of person to create the software you need, for example. Tools are quickly becoming available that allow business creatives to implement their solutions themselves, instead of relying on skilled technical staff.

This alleviates the pressure to find more skilled people and at the same time increases the effectiveness of the company. A true win-win situation, that only requires managers to think out of the box.

With the skills shortage as high as it is now, companies simply cannot afford to keep looking for more and faster horses. They need to find cars.

Automate your manufacturing process!

If your company manufactures goods or products, it will have a certain process in place to do so. However, ensuring that every step of this process goes efficiently and smoothly isn’t easy.

An inventory that isn’t up to date, for example, can cause serious problems – up to the point where you need to stop production until missing parts arrive. Oftentimes, automation provides an excellent solution, especially if you use a tool like Triggre. In this blog, we’d like to share an example which illustrates this perfectly.

Inventory management and up-to-date information

One of our customers in the automotive industry recently asked us to create the concept version of an application within two weeks. After that, they would continue developing it independently.

The goal of the application: automate inventory management based on production, so car parts are always in stock. Obviously, this greatly facilitates the car manufacturing process, as it optimizes efficiency and eliminates the hassle of having to stop production due to a missing part.

But we took it one step further by mapping the individual process steps, based on which we added information to the application. It is now possible to indicate the production stage at which each car part is used.

As a result, employees have an instant overview at every stage. They get to see a list containing all the required parts, and once they use these, the application automatically updates the inventory.

This way, everyone within the organization always gets topical information on the status quo. Moreover, employees make less mistakes during the manufacturing process, as they always know which parts they should use at what time.

Cost-efficient and tailor-made solution

Although deadlines were tight, we met them all, and the application was up and running within the set time frame. More importantly, the results are satisfying.

Especially when running a smaller business – like our customer’s – major ERP systems are rarely money well spent. Therefore, a stable solution such as Triggre provides an excellent solution: it’s cost efficient and entirely tailored to your process!

Pulling back the curtain: a behind-the-scenes look at Triggre

A little while ago, I test drove a Tesla out of curiosity whether it would live up to the hype. In my eyes, it wasn’t all that special; it’s a pleasant and luxurious car, which I’d only expect for the price.

My interest was much more piqued by the almost bare undercarriage that was in the showroom. You could see where the batteries are located and it gave a sense of how the electro engines work. Since I’m probably not the only one who likes to see how things work underneath the hood, I will pull back the curtain on Triggre a bit today.

As previously discussed, Triggre consists of three parts: the Designer, the Builder and the application itself. The Designer of Triggre has about one million lines of code dedicated to only one thing: making designing an application as easy as possible for the user.

After users design their application, they hit the ‘publish’ button. That’s when Triggre-engine really starts running, because approximately thirty seconds later, the application the user designed is up and running in the cloud. So what happens in those thirty seconds?

From designer to up-and-running application

After the user hits ‘publish,’ the entire design is checked once more for inconsistencies in about a second. These can occur, for example, when you have deleted a page and you can no longer reach another page in your application.

Once checked, the design – including all forms, pages, buttons, and processes to be automated – is sent to one of our builders, where it is placed in a queue. Since we have several builders, however, it hardly ever happens that you actually need to wait.

The builder then writes the Javascript, .Net code for the backend, and SQL scripts to set up or migrate the database – briefly put, it writes the application. As our builder can write thousands of lines of code per second, programming the entire application is usually just a matter of seconds.

Installation packages

Like many applications, the code is compiled into a program and an installation package is created. If you are unfamiliar with web applications, this installation package is not completely unlike how you can download a program from the internet and install it on your computer. That too is an installation package of sorts.

Unlike those installation packages, the installation packages of your Triggre application need to run in the cloud, where it automatically is installed besides your already running application.

Once the database migration is completed using the created SQL scripts, the old application goes offline, and the new one goes online. This entire process takes up approximately thirty seconds, after which you can start using your new application!

What then happens, occasionally shocks some people: as the builder cleans up some files, the entire source code of the program just created is thrown away.

For many companies, the source code of a program is a very important commodity and the fact that we simply discard it is foreign to a lot of people. What however is the worth of keeping it, when it can be completely rewritten in a matter of seconds?

Creating an application in seconds, weeks, or months?

What our builder accomplishes in a mere thirty seconds takes the average IT company’s programming team weeks or months. Adaptations to an existing application would require another weeks-to-months-long process, whereas Triggre can implement these within another thirty seconds after your hit ‘publish!’

When digitization goes too far

Many schools use student information systems, and have been for a long time. They store grades and other information digitally. And that is a good thing, as it makes access to these data much easier. However, as I discussed before, such technology also has a downside.

Easy access

While easy access to data can be great, the question is: access for who? In the case of student information systems, this question was never asked. It was simply assumed that access was a good thing.

Since the introduction of the first student information systems, smart phones have made their introduction. Nowadays almost everyone has a smart phone – including parents.

So logical step for student information system suppliers were mobile apps. With these apps parents have easy access to the student information about their children, such as grades and attendance.

By law parents are allowed to see such information, but only until their children reach the age of 16. After that, children need to give consent to their parents to access this information. After the age of 18, parents are not allowed to see this information at all.

Student information system suppliers however, provide information to parents without any consent ever given by students themselves.


And while this is strictly a legal issue, there is a more important factor to consider. One parent said that such access takes away a particular kind of joy from the student.

Imagine the student getting a really good grade. If the parent immediately knows this grade, because they receive a push notification, the student is robbed of being able to come home and telling their parents about their grade.

Even with bad grades, shouldn’t it be up to the students to decide when to tell their parents? Isn’t that part of growing up? And shouldn’t student information system suppliers know this, and even support this process?

The goal

I believe this is a perfect example of taking digitization too far. While the technology is there, it doesn’t have to be used. If the goal of parents is to raise their children to be responsible adults, using technology to effectively spy on their children teaches the wrong lesson.

This is something that happens too often. Companies do not fully understand the goal at play. They cater to the needs of their customers, while often the true goal lies one step further.

In the case of student information systems, the goal is not the school or the parent, but teaching the student – that’s the real goal.

Maintenance on software is just like car maintenance: you’re getting played

In 2008, I brought my car to the shop for a periodic check-up, because it had a problem. Sadly, the check-up showed my car was in desperate need of repairs. The mechanics told me that it was probably not worth the cost.

It was a very, very old car and I had seen this coming for a while. Still, a mechanic telling me I’d best buy a new car, wasn’t really what I had in mind when I brought it in.

Adhering to a strict schedule

My first car was an old second-hand one. To be honest, I hadn’t been too strict with the maintenance windows. So, when I had to suddenly replace it, I vowed that when I would buy a new car I’d always take it to the shop on time, to make sure it was always in perfect condition.

The upside of taking your car to the shop on a regular schedule is that it is indeed always in great condition. The downside however: you get told by mechanics that something needs to be changed and you basically don’t know whether it is strictly necessary.

So, you have them replace it. When it comes down to it, this means spending a lot of money on things that do not seem broken.

Maintenance on custom built software

Any application that is custom built, or has a custom built component, will require maintenance. The simple fact of the matter is that the world around that application or component changes, which may require the application or component to change in return.

This is actually quite similar to owning a car. You buy the car, and because you use it, it requires maintenance. Just like software requires maintenance, due to the world in which it is used changes.

You bring your car to the shop for a check-up, the same way your software requires some maintenance every once in a while too. All to keep things running smoothly. For custom-built software however, most companies do not perform maintenance as often as is required.

Instead, they prefer the method to use it for as long as humanly possible, after which the whole thing needs to be replaced. Much sooner, though perhaps at slightly lower cost, than performing periodic maintenance.

Knowledge gap

If it isn’t broken, don’t fix it. It seems like sound advice. But if you don’t, it means using everything until it is absolutely worn down and replacement becomes paramount.

There is a huge gap between following every advice from a car mechanic and only replacing parts when they’re totally worn out.

This gap is a knowledge gap. If you’re like me, you don’t know everything about the car you own. This puts me in a disadvantaged position whenever a mechanic tells me something needs maintenance or replacement, especially if it isn’t causing a noticeable problem.

If I trust the mechanic, I’ll let them replace it. If I don’t, it is usually about time to bring my car in to another shop.

Bringing your software to another shop

In the past, I have switched shops for car maintenance a few times. This is fairly simple, since cars are a common thing and there are many shops that can service my needs. If I am unsatisfied, I simply take my business to a competitor.

And for software it’s pretty much the same deal, right? Software is very common. There are many companies that can build software and maintain it. So if I am unsatisfied, I can just switch to a competitor, just like I did with the car. Or can I?

The problem is that software, especially when it’s custom built, requires an understanding of how it works. Most applications also communicate with other applications, and are implemented to support a business process.

This makes understanding ‘how it works’ a lot harder. So, unlike bringing your car to another shop, bringing your software to another ‘shop’ would cost a lot of time and money.

The new shop would first need to get a good understanding of your business process, the other applications that the software communicates with and then understand the code of the application itself.

No leverage

Inevitably, you are left with without any leverage. Because if you cannot bring your software to another company easily, you only really have 3 options:

1)      Don’t perform any maintenance at all
2)      Just accept anything that your supplier tells you should be done
3)      Insource maintenance of the application

We believe that this status-quo should change. Not performing maintenance means losing innovative strength, since adapting software to changing processes or markets is no longer an option.

Accepting anything that your supplier sells you will cost too much money for what you are getting in return. And lastly, insourcing maintenance is extremely costly, if not impossible, due to the shortage of IT professionals.


Triggre strives to give companies back their freedom when it comes to implementing and maintaining software. Just having the option to make software yourself (without IT specialists) is a game changer, which makes you far less dependent on external companies – including us.

Why your kids shouldn’t pursue a career in coding

When I was in high-school, in the nineties, there was a simple expectation that my mom had. I was to get good grades, so I could go to university. Once in university, I would graduate in a respectable time frame. With my university degree I would be ensured a good job.

My mother’s generation, born in the 50s and 60s, all raised their children this way. And that made sense, because the people with the best paying jobs were notaries, doctors and lawyers.

Notaries, doctors, lawyers and pilots are all last-century jobs

The glory days of notaries, doctors, lawyers and other similar jobs are far behind us. Even being a pilot, once the pinnacle of achievements, is long past its golden era. It may not always look like it, so bear with me. There are still people who make a lot of money in those last-century jobs. Usually they have been in the job for a long time.

Take pilots for example. Pilots who are over about 45 years of age, still earn a lot of money. The younger pilots however, have a very rough time. Demands to get a job have increased significantly, such as the amount of hours flown, which has doubled in recent years for the same position. This means that pilots have to accept jobs in other countries, where safety regulations are less, pay is much lower and the hours are crazy. And that is a big problem, because becoming a pilot is very expensive. Most young pilots have between USD 100.000 and 200.000 in debt, which they can hardly pay back with their relatively small wage.

Notaries have a similar hard time. Where the hourly rates used to be very high, now the rates are plummeting. There are simply way too many notaries and the work that they do can very easily be automated. For example, Dutch company HEMA introduced an online notary last year. Most documents you get are standard, with only a small number of possible options. Now, people can get such documents at a very, very low price, online.

Automation is destroying last-century jobs

Last-century jobs are simply not paying as much because there are too many people who do that same job. And even if there isn’t an abundance, last-century jobs are subject to automation. When part of the job is replaced by automation, there is less work for the remaining work force. This means there is more competition. And as we have all learned, if there is more supply than demand in a competitive market, prices will drop. In this case, that means wages and hourly rates. Ultimately, this is what costs people their job.

Automation isn’t a bad thing though. It has been going on as long as mankind has been working. The pyramids were built using huge amounts of manual labor, which is very likely the reason there aren’t that many of them. Today, we use automated muscle to build sky scrapers. That automation has significantly increased the number of buildings we can produce, hence the skylines we see all across the world.

And while automation may have cost us purely physical jobs such as construction work, it has freed up time for other jobs that could not be automated by machinery. Doctors, lawyers, notaries and pilots for example … Those jobs are now facing the same fate as physical jobs before them did. As their jobs are automated more and more, prices and wages will drop.

If you’re a millennial, you wouldn’t think of advising your children to become a lawyer or doctor. The world is different now and what was once a guaranteed high-paying job, is now just another job impending automation.

Coding will be automated as well

If you take any lesson from history, it is that what is now a high-paying job is very likely to be automated. Think about it. If something costs a lot of money, it is worth automating. Especially if there is a lot of need for it. In my previous blog post, I gave some numbers about the shortage of IT specialists. That shortage is quickly giving rise to many automated solutions that will replace coders.

For the foreseeable future, coders will still be needed. I think coders will be needed for at least another 50 years. The question is, how many? Right now, we need a lot of them. We’re building the software equivalent of pyramids. In the meantime, we are automating parts of what is now done by coders. The combination of automation (which decreases demand) and the attractiveness of becoming a coder due to very high wages and rates, will soon tip over. My estimation is that within 10 years rates for coders will drop significantly.

In about 20 years coding will be as polarized as other jobs in danger of automation are today. You are either extremely good at your work, or you work for such a low wage that it’s cheaper than automation. It’s not a good position to be in honestly.

Teach your kids creativity and human connection

If you are a millennial, you don’t want to make the same mistake my mom’s generation made. You don’t want to advise your kids to choose a certain job that, if you had known how technology will advance, will be obsolete in a few years. You don’t want them to pursue a career in coding.

Technology is advancing in a direction where we are now automating minds. Computers are now good at doing things doctors were once good in. In some areas, computers are even better than humans. One example is recognizing tumors on mammograms, where computers clearly outperform doctors. Similar advances are seen in other areas, such as self-driving cars. Some may take a long time to become better than humans, but they will definitely be better than humans someday in the near future.

If such jobs, that once were the uncontested domain of humans, are automated, what is the next thing? In my opinion we are still very far away from computers replacing true human connection and creativity. There are plenty of examples of artificial intelligent machines knocking on the door of creativity, such as writing music or making paintings. All such examples are however, based on huge numbers of pre-existing information. Artificial intelligence can use millions of musical compositions to learn what we think is nice music. But it will not invent jazz, grunge or dance music in a world where it doesn’t exist yet. That requires true creativity.

Human connection is another thing that is very hard to replace. Think about it. We don’t even truly know how human connection works. We know a lot about it, but it isn’t an exact science. In a world where everything will be automated, easily accessible and done by machines, human connection will become a very valuable asset.

So if you have to decide what you want your children to be good at, think about this. All kids play with computers, fewer and fewer play with each other. Many kids know how to use things, but fewer and fewer kids know how to create things. Give your kids creative toys such as Lego and teach them to interact with other kids by letting them play together. When they get older, keep feeding their creativity. Let them take up things such as photography, painting, mechanical engineering and coding. The more different experiences, the better. That’s what ultimately leads to creativity. As long as you understand that each thing separately will be automated. It’s the combination of creativity and human connection that are going to be the valuable skills later this century.