Artificial intelligence is changing how IT teams create, implement and support digital solutions. For Ukrainian companies, this creates opportunities to develop their own products, improve service quality and solve customers’ problems faster. Making use of those opportunities requires the ability to combine engineering, an understanding of the customer, and responsibility for the outcome.
These changes concern the entire IT sector: product and service companies, systems integrators and in-house IT teams. AI tools can assist with programming, testing, data analysis, documentation and user support. In tasks where they reduce routine work, companies need to reconsider how they organise that work and what makes their services valuable.
For the Ukrainian IT sector, this is a question of company strategy, professional development and what customers will be willing to pay for in the years ahead.
When a task can be completed faster, the amount of effort it previously required is no longer sufficient to justify its price. A customer can reasonably expect shorter delivery times, a more affordable solution or more useful work within the same budget. A company needs to explain its contribution more precisely: the problem it solves, the risks it takes on and the outcome it delivers.
At the same time, cheaper development does not automatically create demand. A team can quickly launch a digital service for which there are no customers. It can build an enterprise application in a few days that merely gives employees another system to fill in. It can automate a process that should have been simplified first.
The speed of an IT team’s work makes economic sense when it shortens the path to a useful outcome.
A document approval system provides a useful example. Generating forms, workflows and notifications may take progressively less time. But the customer needs to shorten approval cycles, reduce errors and see who is responsible for each decision. This requires an understanding of the actual process, its exceptions, access permissions and interaction with other systems.
This is where product and engineering expertise come together. Someone who understands only the desired interface may miss a critical constraint. Someone who sees only the technical specification may implement a flawed process very well. The team needs to be able to examine both sides.
That is why I am cautious about forecasts in which product managers automatically become the main winners while programmers lose their relevance. A job title does not guarantee an ability to create value. A technical specialist may be the first to identify a way to solve the problem. A product manager may spend months on features for which demand exists only in a presentation.
AI can quickly increase the number of proposed ideas and features. Without clear priorities, this creates more switching between tasks and increases the amount of unfinished work for the team.
A good product decision involves making a choice. The team needs to decide which problem to solve now, which opportunities to set aside and what evidence will count as success. Knowing when to stop an unsuccessful initiative can save more resources than faster coding.
A finding from DORA: AI-assisted Software Development (2025) helps explain this effect: AI amplifies an organisation’s existing strengths and weaknesses. If a company can set priorities, check quality and act on feedback, new tools help it. If these processes are weak, faster production can increase the amount of rework.
Meanwhile, engineering complexity does not disappear at a manager’s instruction. In an enterprise system, technical decisions affect data integrity, security and the ability to deploy updates without interrupting the business. In a cloud service, they affect stability under load, infrastructure costs and recovery after a failure. In an integration project, they determine whether systems exchange data correctly. All of this requires engineering work throughout the solution’s lifecycle.
A new feature can be compared with an optimisation only in the context of a specific need. If an optimisation allows an application to run on the customer’s hardware or significantly reduces infrastructure costs, it has direct product value. An attractive feature that nobody uses may have none.
Research is already identifying meaningful improvements in the speed of certain kinds of work. The study Microsoft Research: AI and Developer Productivity (2025) combines three field experiments involving 4,867 developers. Access to an AI assistant produced a combined increase of approximately 26% in the number of completed tasks. This estimate applies to the developer teams studied; the effect on other tasks and roles needs to be measured separately. However, that metric does not yet answer questions about the product’s long-term quality, adoption or profitability.
For a manager, this means examining the entire delivery cycle. How much time passed between identifying a need and meeting it? How many changes had to be reworked? Did the customer experience fewer errors? Did users start completing their work faster? The amount of generated code does not answer these questions.
For Ukrainian companies, I see an opportunity to deepen their industry expertise. Knowledge of manufacturing, logistics, energy or public administration processes helps identify which problem is worth automating. AI can accelerate the creation of a solution, while expertise can improve the likelihood that the solution will be needed.
For a product company, this means creating a solution that people use and are willing to pay for. For a service business, it means helping the client move from requirements to implementation and reliable operation. For an in-house IT team, it means ensuring process continuity, reliable data and tools that employees can use effectively. In each case, value needs to be tied to a specific outcome for the user or the business.
This also gives companies a reason to reconsider their commercial offering. A company can combine development with process analysis, integration, support and measurement of the impact. Outcome-based pricing requires clear agreements and a division of responsibility, but even under an hourly billing model it is useful to show the customer the connection between the work delivered and changes in their business.
Employers have a separate responsibility to develop junior engineers. Automating simple tasks changes the familiar learning path. Companies need to give newcomers opportunities to understand solutions, verify generated code and deal with the consequences of their own changes. Otherwise, today’s savings on training could become tomorrow’s shortage of people capable of evaluating complex systems.
For a specialist, the best response to these changes is to broaden their expertise: master AI tools, maintain engineering discipline, understand product economics and talk to users.
The next step for the Ukrainian IT sector is to learn how to move faster through the entire journey from a need to a verified outcome: in its own products, in client projects and in the day-to-day operation of enterprise systems. The ability to complete that journey, ensure quality and take responsibility for the consequences gives the customer a reason to pay.
The Association’s work in this area: Software Ukraine AI Commission.