Anyone who has been involved in an IT project at least once knows that the first version of a cost estimate usually lands in the trash. The reason? Underestimation. The premise was beautiful: the system was supposed to be simple, fast and cheap, but you end up with delays, nerves and surcharges.
Implementing IT systems can be like building a house. We start with a vision: “We want a nice, modern house of 120 meters, and a large terrace to go with it.” But only then do we discover that foundations, installations, connections are needed. Finishing costs start to overwhelm us. It works the same way in IT – only that the hidden costs of AI implementation can reach half the budget. Let’s see where the financial pitfalls lurk.
If someone were to ask me about the biggest paradox of implementations, I would answer without hesitation that the main contender for this title would be the temptation to save money on requirements analysis. The shortcut may seem attractive, because creating good documentation involves weeks of meetings, workshops and analyst work. However, it can yield uninteresting results, because skipping any of the important preparatory steps can result in a project that takes off with false assumptions and ends up with revisions that are more expensive than a solid start.
Why invest in production automation when we have a much cheaper manual system? Let’s take a look at whether it’s certainly cheaper. Every data error – a misspelled invoice number, a mistake in machine parameters, a misrecorded machine time – is a cost that comes back like a boomerang. Automation requires more investment at the beginning, but minimizes the risk of human error.
Imagine an operator who, at the end of a night shift, has to read a series of digits from a machine or give a breakdown instruction to a shifter. Fatigue does its job and it’s easy to make a mistake. And a machine’s downtime counted in hours can consume more than a year’s license of an AI-based diagnostic support solution.
“We can handle ourselves” – I can’t count how many times I’ve heard this phrase. The problem is that without experience, it’s easy to duplicate the mistakes of other companies. Consultants cost money – that’s true. But their role is to shorten the path to implementing IT solutions and protect you from costly slip-ups. If you don’t opt for professional technology consulting it’s like trying to overhaul a car based on YouTube videos. It can be done, but are you sure you want to take the risk?
When it comes to the cost of implementing IT and AI systems, there is no single ideal solution. The cloud can sometimes be cheaper and more flexible, but sensitive data often has to remain on local servers. More often than not, you end up with a hybrid – and this is where additional integration costs come in.
When you buy a suit, you want it to lie perfectly on you. It’s hardly surprising. It’s no different with software and IT systems. Companies want the solutions they implement to “fit them like a tailor-made suit.” The problem is that such a suit is expensive to maintain. Every update or change requires additional work. Sometimes it is better to fit the process to a proven platform than to bend the software by force.
Get 5 chapters of the book for free!
Join the newsletter and gain access to 40% of the book
“15 Steps to Buying an Information System“.
When buying a system, we often look only at the price of implementation. Meanwhile, annual license fees can exceed the initial cost within a few years. That’s not all. You have to reckon with other limitations such as the number of users. This is because often each additional login means another invoice.
Servers, operating systems, databases – who will give more! Many managers are tearing their hair out, because, after all, it wasn’t in the cost estimate, and meanwhile without all this the system won’t start. It’s like buying a luxury car and only upon delivery finding out that the wheels need to be bought separately.
You have invested in a new IT or AI system. Fascinated by the new solution, you come up with a great idea to expand it with new functionality. Well, the appetite grows as you eat, and new ideas know no limits. “Since we are already building a system, why don’t we add a module for logistics,” you think. The problem is that each additional module means a budget and deadline shift.
Let’s go back to the construction of the house. Let’s assume that the walls are standing, the lawn is seeded, the lamps are screwed in, paintings are already hanging on the walls, and you’ve even put flowers on the windowsill. After a month, the faucet broke, and after four you thawed the tiles on the terrace. The house – a bottomless well – now you know what your friends were talking about. An IT system is not a one-time purchase either. It needs updates, backups, monitoring and people to take care of it. Many companies assume that updates are a job for the internal IT department. But when the first failure comes, they find they need additional expertise or a maintenance contract.
Even the best system won’t work if people don’t know how to use it. Training, instruction, user support – this is a cost that is often downplayed. And yet, any change in a company also involves emotions: resistance, stress, a sense of danger. AI or MES systems are no exception. Therefore, think about how to better manage change.
There are three ways to minimize the risk of disappointment due to hidden costs of IT and AI implementation:
Hidden costs are not the enemy of the project. They are a natural part of it. Companies that include these additional expenses in the budget not only avoid disappointment, but most importantly achieve a faster return on investment. Because implementing an information system or artificial intelligence is not a sprint for a cheap solution. It’s a marathon in which the winner is the one who can anticipate the turns and spread the forces.
How about your company? When preparing for digital transformation, does it look only at the first page of the cost estimate, or at what’s written in the fine print at the end?
