The widespread adoption of artificial intelligence by Ukrainian businesses has led not only to process optimization but also to significant technological and managerial challenges. This evolution is detailed by Dengi.ua in their recent report .
For major entities like Epicenter and Vodafone Ukraine, the primary barrier has been integrating algorithms into legacy IT infrastructure. Since corporate systems were not originally designed for neural networks, these companies are now forced to restructure their entire data architecture.
MHP Agroholding confirms this hurdle, noting that any AI initiative faces a fundamental challenge: data maturity. Specialists emphasize that AI cannot function effectively without robust data infrastructure, clear process logic, and strict "data discipline."
Uklon, the ride-hailing service, has also encountered technological obstacles. According to CTO Oleksandr Chumak, initial AI models struggled with context volume limitations, causing agents to "lose" the thread of a conversation. The company resolved this by transitioning to a multi-agent architecture featuring highly specialized bots.
To mitigate "hallucinations" (the generation of false facts), Uklon implemented mandatory source citation and daily automated quality audits. Furthermore, the company had to manage token consumption, as automated scripts occasionally processed extraneous data, leading to unnecessary costs. Consequently, Uklon shifted from flexible local models to more cost-effective and accurate off-the-shelf enterprise solutions.
Security remains a paramount concern. To prevent sensitive data leaks, Vodafone deploys most AI solutions locally using its own GPU clusters. Similarly, monobank avoids sharing data with external providers, choosing instead to develop proprietary, closed-loop models and platforms.
Meanwhile, Ribas Hotels Group emphasizes that the primary obstacle is often cultural rather than technical. Skepticism among employees and management regarding AI’s impact on their workflows frequently acts as a brake on innovation.