The expert named the main mistakes of companies in digital transformation
Russian industrial enterprises and fuel and energy companies are accelerating digital transformation. However, the introduction of new solutions is often accompanied by mistakes, from inefficient import substitution to overloading of IT teams. On August 14, Dmitry Pilipenko, Deputy CEO of LANSOFT, told Izvestia which miscalculations cost businesses the most and how to avoid them.
The first mistake is to ignore the fact that any fashionable technology has its limitations and disadvantages. This often leads to the fact that all tasks are solved using fashionable technologies, even if they are not optimal in this particular case. For example, microservice architecture has a lot of advantages, but its use in ERP greatly increases the volume of integrations between microservices, which are very difficult to manage.
"The fate of low-code is also indicative. A year ago, it was widely considered as a universal tool for rapid digitalization, but in 2026, interest in it has noticeably cooled: according to market estimates, the share of companies that see it as an independent value has decreased by about half. The reason is not that the technology is bad, but that they tried to use it in isolation from the architecture — where it did not scale, did not integrate well, or simply did not integrate into real processes," the expert said.
The second mistake is to implement solutions based on artificial intelligence (AI) in pursuit of fashion, without having a sufficient amount of structured and cleaned data. This is especially important for industrial enterprises. Often, the necessary parameters are simply not measured. The next level is that the sensors are installed, but the data storage depth is insufficient. Finally, model training requires not just raw data, but cleaned, structured, and labeled data. All this requires separate projects, capital investments and skilled labor, which will be discussed below.
The third mistake, according to Pilipenko, is to consider regulatory requirements for import substitution as an annoying burden when their implementation is reduced to the formal replacement of a foreign product with a domestic equivalent from the registry — without changing the processes and architecture. This leads to the hasty implementation of insufficiently mature solutions.
"A consistent strategic approach based on several important considerations would be more appropriate. First, the main thing should be to ensure business continuity: new solutions should not fail either in terms of reliability, convenience, or functionality. Secondly, import substitution should be used as an opportunity to switch to modern technologies and finally abandon old, not always effective, and often inherited solutions. Finally, it is possible to redesign the functional appearance of the systems and include those business requirements that have been in the backlog for years," the specialist emphasized.
The fourth mistake is to continue to implement all planned projects when reducing IT budgets, but either at a slower pace or with an overload of available resources, or more often both at once.
Pilipenko called prioritization of projects a reasonable approach. First, business-critical projects should receive the highest status, and secondly, they should be strategically important, even if they don't pay off quickly. For example, equipping sensors, deploying communication networks, or implementing a data management platform.
Trying to pull 10 projects at the same time leads to the fact that none of them is completed on time, and the team burns out. Less, but it's almost always more effective to the end than a lot and half.
The expert added that the fifth mistake follows from one of the previous ones: to assume that AI agents completely replace IT personnel, and you can increase the volume of work without looking. In fact, using agents for programming does not reduce, but rather increases the need for qualified and rare specialists. Firstly, these are architects who are able to conduct an orchestra of agents, resulting in a well-built product. Secondly, these are industry specialists who deeply understand the possibilities of IT technologies and are able to set the task for agents.
Pilipenko concluded that a business that is entering into a complex implementation should solve the personnel issue in advance: develop internal expertise, invest in the growth of its people and retain key specialists, and not expect that the right person will be able to hire quickly from the market.
On July 7, State Duma deputies at a plenary session adopted in the second and third readings a government bill to support the development of artificial intelligence technologies. In an explanatory note, the parliamentarians pointed out that the document is aimed at creating legal conditions for the accelerated development and implementation of large fundamental AI models in the Russian Federation.
Переведено сервисом «Яндекс Переводчик»