You may have digital colleagues: what Ukrainian IT top managers expect from 2026

One of the things that inspired IT professionals this year was the huge leap forward in generative video. According to Dmytro Lola, CEO of Kiss My Apps, high-quality video can now be produced in a matter of minutes: "This is changing the industry, including creative production."For more on other trends and expectations for 2026, according to Epam, SoftServe, Genesis, GlobalLogic, Intellias, Sombra, and Kiss My Apps, check out this article on LIGA.net.The main trend for 2025 – agentic AIAll companies surveyed named agentic AI as the defining trend for 2025.There has been a shift from "request-response" systems to systems that can operate autonomously, coordinate tasks, and make decisions without constant human supervision," says Pavlo Ryabchuk, head of the Digital Solution Office department at Intellias, describing the changes in the industry over the past year.To make this autonomy effective and safe, the industry relies on two related technologies: Edge AI (running AI algorithms directly on local devices) and Enterprise AI (using AI across an entire company). According to Ryabchuk, these are the driving forces for the technology industry in the near future.Yuriy Nakonechny, CTO at Sombra, says that 2025 has become the year of practical commercialization of AI agents and the transition from experiments to real business cases. "The tech industry was driven by tools that accelerate SDLC, reduce time-to-market, and put engineers into a mode of working ‘with models’ rather than ‘instead of models,’" he explains.According to Oleg Tsal-Tsalko, CTO of Epam Ukraine, the pace of the year was set by the battle between OpenAI, Anthropic, Google, Meta, AWS, and Chinese players, as well as the fact that AWS, with its new Graviton5 and Trainium3 chips, imposed competition on Nvidia's hardware.Among programming languages, Python plays a leading role, also thanks to the spread of AI. Rust is also gaining popularity due to its memory model and performance, especially in mission-critical production systems.How companies use agentic AIDmytro Lola, CEO and founder of the Kiss My Apps platform company, notes that this year Kiss My Apps implemented AI personalization and on-device machine learning. They use local neural models in their applications, which allow them to process data in real time without relying on cloud services.Epam is building and developing EPAM AI/Run Agentic Platform, a platform that allows you to create teams of agents. Head Office Genesis has automated processes in the Compliance team: now one specialist saves at least 60 hours per month. And in 2025, Global Logic continued to develop VelocityAI, an ecosystem for implementing AI in SDLC, developed jointly with Hitachi:"20+ agents help more than 33,000 specialists in key functions (from preparing technical tasks to sorting incidents)", according to the company. "For example, today, some of the routine checks are performed not by testers, but by an automated system. This has allowed us to optimize processes by 20-40%."At the same time, all speakers caution that AI is not yet ready to be either a universal answer to everything or a complete replacement for specialists.For example, at SoftServe, fully autonomous AI teams are considered hype because the technology is not yet "mature enough." In their opinion, the greatest value now comes from the symbiosis of engineer and agent. GlobalLogic expects a shift from prompt engineering to context engineering in the future, which will accordingly affect the market and engineering requirements.7 trends for 2025In addition to AI, the companies identified the following trends for 2025:Zero-trust architectures. An approach to cybersecurity where every request, device, and user undergoes a new access check.Hybrid computing. Combining local infrastructure, cloud, and edge solutions. Businesses are moving to hybrid architectures to optimize performance, cost, and sensitive data handling.Low-code/no-code. Platforms that allow applications to be built without traditional programming. Accelerates development and reduces the workload on engineering teams.Data-driven enterprise. The transition to a model where every management or product decision is based on data. Companies are investing in data infrastructure, data quality, and analytics.Multi-platform. User expectations of a consistent experience across all devices are forcing companies to build products that work on all devices with a single logic and backend.Regulation and Responsible AI. Tighter regulations require companies to implement ethical and secure AI. In 2025, this became a separate competence within large teams.Competition is moving into the hardware layer. The battle is taking place not only at the model level, but also at the level of GPUs, specialized chips, data centers, and power grids. Companies are looking for ways to make models cheaper to compute, rather than endlessly increasing their size.What inspired them in 2025"We were most inspired by AI-based developments that have a real impact on humanity and the planet," says Oleg Tsal-Tsalco from Epam. "They help combat deforestation, analyze ecosystems, save coral reefs, and predict climate risks."Tsal-Tsalco was also inspired by cases of AI use in medicine, the public sector, and especially defense tech.Pavlo Ryabchuk from Intellias noted that he was most interested in observing innovations that "transform abstract AI into a working engineering tool." Genesis was inspired by developments that allow the team to work faster and with fewer errors, as well as technologies that change the business model.Yuriy Nakonechny from Sombra says he is most impressed by solutions that integrate agents directly into development: from RFP analysis and architecture generation to automatic unit test writing and CI/CD optimization.Dmytro Lola from Kiss My Apps was most inspired by Sora and Nano Banana."Generative video has made a huge leap in a year," says Lola. "It is changing the industry, including our creative production. The process is becoming faster and cheaper for both product content and advertising."According to Lola, Sora 2 can generate videos with sound in a single request, "literally in a few minutes, without editing." Google's Nano Banana allows you to edit photos with text prompts, such as "add sunlight."Expectations for 2026"2026 looks like the year when many companies may see the emergence of digital colleagues who handle individual processes or parts of them (support, analytics, operations)," says Genesis.The company also believes that the share of domain-specific models will grow.The more AI appears, the more risks arise. Therefore, among the trends for 2026 will be the creation of standards, regulation, and new security solutions, as mentioned by several companies surveyed, including Sombra.At SoftServe, where AI is showing 85% annual growth, they see three main paths for the development of artificial intelligence in 2026:Physical AI – the synergy of generative AI and hardware.Multi-agent systems – where not one universal AI performs tasks, but dozens of specialized agents collaborate with each other.Multimodal AI – which can see different data formats and combine them into a single coherent context. SoftServe has already developed the Multimodal RAG solution together with Nvidia, which simultaneously analyzes text and images (including tables and diagrams). The company says that this approach increases the accuracy of results by more than 70% and reduces the time spent searching for information by approximately 40%.According to Epam, robotics and robotics will receive a new impetus, as will the introduction of AI in govtech. The company also expects the first breakthroughs in quantum computing.We may also see a breakthrough in EnergyTech, according to Intellias. Here, progress is expected in the field of thermonuclear energy and the development of solar photovoltaic (PV) technologies. In addition, according to Intellias, special attention will be paid to user interface (UI) personalization next year.Kiss My Apps also talks about personalization, or rather hyper-personalization, particularly through AR and VR technologies. They predicts that by 2027, AI models will shrink so much that calculations can be performed on devices rather than servers. The transition to this will begin as early as 2026.Kiss My Apps also expects preparations for 6G to begin in 2026. "Although commercial use is planned for 2030, active deployment of infrastructure and laying the foundation will begin as early as 2026," concludes Dmitry Lola.