According to a 2026 AI usage study by Török Balázs, an AI expert and digital marketing consultant, Hungarian employees are increasingly integrating artificial intelligence into their workflows, but the advantages that come with AI — such as access to training and time savings — are not evenly distributed across groups.
What the study examined
The survey assessed how AI tool usage changed compared with 2025, how much weekly time respondents save, the role of AI training, the adoption of AI agents and Custom GPTs, and data security issues related to using sensitive personal or confidential company data with AI tools.
More training, but many still use AI without organizational support
One key finding is that the proportion of employees receiving workplace AI training increased across all company sizes. The largest rise appeared at companies with more than 250 employees: while 45 percent of respondents in this group reported AI training in 2025, that share grew to 68 percent in 2026. In medium-sized firms the rate rose from 27 to 45 percent, and in companies with 1–49 employees it increased from 31 to 37 percent.
Despite this progress, the study highlights that many employees still use AI tools without organized support, often relying on personal accounts and devices. This is especially challenging for smaller firms, which typically have fewer resources for internal training, tool procurement and establishing AI usage policies.
Training correlates with greater efficiency gains
Responses show a positive link between participation in AI training and the amount of time saved at work: employees who both have access to AI tools and received training are more likely to convert the technology into measurable efficiency gains. Sárospataki Albert, co-founder and CEO of Billingo Technologies Zrt., said well-applied AI can free up time, support better decisions and raise productivity.
AI agents, Custom GPTs and multi-platform use bring larger time wins
The study found that respondents who use AI agents or Custom GPTs report significant weekly time savings more often — including frequent reports of 10–40 hours and sometimes more than 40 hours per week — compared with those who only use simpler prompts or occasional AI solutions. This suggests AI is increasingly used to support whole workflows rather than just speeding up individual tasks.
ChatGPT remains a dominant tool among Hungarian workers, but multi-platform usage is strengthening: respondents use different platforms for different tasks, such as ChatGPT, Gemini, Claude, Copilot or other AI solutions.
The survey also noted substantial efficiency gains among sole proprietors, even though they generally operate with less organizational support and fewer resources. Mórász Márton, founder and CEO of Proactive Business Zrt., observed that time savings achieved by sole proprietors are approaching the level of senior executives, which directly translates into more revenue or free time for them.
Data security, access control and AI governance are key issues
A focal topic of the research was the use of sensitive personal and confidential corporate data in AI tools. Those who incorporate such information into AI-enabled workflows report larger weekly time savings, but this raises serious data protection, access control and client-trust questions.
The mixing of personal subscriptions and corporate access is particularly sensitive: if employees work with company data using personal accounts or uncontrolled devices, it may improve short-term efficiency but creates long-term organizational risk. Varga János Tamás, founder and office-managing partner of VJT-Partners, argues that AI use should not be banned but framed: clear rules, controlled tools and conscious AI governance are required to make time gains sustainable.
Implications for companies
Based on the 2026 study, a key task for companies will be to avoid treating AI as isolated, individual experimentation. Genuine competitive advantage requires training, secure tool usage, clear internal policies and a corporate culture that supports innovation while managing data protection risks.
