Paulius Insoda, head of „NFQ Technologies“, warns: blind AI implementation is a direct path to costly mistakes. According to him, most problems today arise not from AI itself, but from how organizations try to apply it in their operations.
Implementing AI because it’s trendy: wanting to keep up with competitors or be innovative
One of the most common mistakes is implementing AI because it’s fashionable, due to market pressure or competitors’ actions, rather than real need. According to P. Insoda, many managers today feel afraid of falling behind, so AI becomes a kind of “must have” element.
“Some companies jump on the AI train as blindly as they did a few years ago on blockchain technology. AI is implemented not because its value is understood, but because shareholders pressure or fear of falling behind competitors gnaws at them. The result? The business buys an expensive toy but has no idea what real problem it should solve,” he says.
According to him, AI itself should not be the goal. This technology creates value only when it solves a specific business task – reducing costs, streamlining processes, improving customer experience, or helping make more accurate decisions. Also where there are large amounts of data, repetitive processes, or high labor costs.
“If a company receives a few customer inquiries per day, automated inquiry processing with AI is probably unnecessary. But if dozens or hundreds of people serve customers and physically cannot keep up with the flow – then the situation is quite different,” explains P. Insoda.
AI is “imposed from above,” and employees start to resist
According to P. Insoda, another common problem is that employees are insufficiently involved in AI implementation, and it is not explained what value the change will create.
“AI initiatives often originate at the management level, but the people who work with the processes daily remain uninvolved. Employees do not understand why it is needed and how it will affect their work, and eventually start to fear that automation will replace them, so they naturally resist changes. We have seen absurd situations where programmers hide AI capabilities from managers who understand technology less well because they feel: if I show how fast I can work, tomorrow either I or my colleagues will become unnecessary,” he says.
According to him, one of the biggest mistakes is thinking that it is enough to implement the technology and people will start using it on their own.
“You need not only technology but also preparation: clear communication, training, habit changes. Otherwise, AI capabilities simply remain unused,” says the interviewee.
Read more Not all «autonomous» cars are autonomous: when to expect a breakthrough in Lithuania?
Automated chaos: implementing AI without organizing data and processes
Another common mistake is trying to implement artificial intelligence in organizations where processes and data are still disorganized. In such cases, AI does not solve the problem but only highlights it.
“AI greatly accelerates all processes, including internal company chaos. If a company does not have systematized data and functioning, not just described, processes, AI will only further expose organizational disorder. AI works best where there is a lot of data and it is of good quality. If such data does not exist, a person often makes better decisions than AI,” says P. Insoda.
P. Insoda tells about a company that wanted to automate an internal process with AI. Technologically the solution worked, but soon problems began in the organization.
“It turned out that the process itself was malfunctioning, unsystematic, and worked differently than described. Without AI, the process somehow worked, challenges arose but less, because everything happened more slowly. When the process was automated, chaos just started happening faster and on a larger scale. AI quickly reveals the weak spots of an organization. If the process does not work without technology, it often will not work with it either – just faster,” he says.
Most managers err thinking that AI will understand and fix process errors by itself. In reality, automation works only where there is clear logic, streamlined processes, and data structure, not unique, “intuitive” work done each time.
Those who do not adapt will be “with a horse while others drive cars”
Although P. Insoda critically assesses some of today’s AI implementation practices, he is convinced that the technology itself will fundamentally change the labor market and business daily life in the coming years.
“AI is not a panacea or a cure for all diseases. It is simply a very powerful tool that can no longer be ignored, because I am sure that those who do not use it properly will still be riding horses while others have long been driving cars. However, it is important before using it to clearly define what you want to change and why, what value it should create. The greatest value in the future will be created not by those who buy new technology the fastest, but by those who can meaningfully apply it in their operations,” says P. Insoda.
