Industry

AI-generated text

AI tools become daily practice for Hungarian developers, saving ~312 hours per year

A 2026 DevHorizon Hungary study commissioned by Amazon Web Services finds 82% of Hungarian software developers use AI tools daily, reporting an average weekly time saving of 6.5 hours (approximately 312 hours per year).

AI tools become daily practice for Hungarian developers, saving ~312 hours per year

AI-based tools are now widely used by Hungarian software developers: the DevHorizon Hungary 2026 survey found that 82% of respondents use AI tools daily and 94% at least weekly. The study was carried out by Strand Partners on behalf of Amazon Web Services (AWS), polling 500 professional software developers working in Hungary between 18 July and 5 August 2026; the sample is representative by sector, company size and region.

How much time does AI free up?

Most respondents reported productivity improvements: 77% said they experienced increased efficiency and 32% described that increase as significant. On average, developers save 6.5 working hours per week by using AI tools, which—assuming 48 working weeks per year—equates to roughly 312 hours saved per developer annually.

Developers allocate the time saved as follows:

  • 33%: solving more complex technical problems
  • 30%: reviewing results produced by AI
  • 27%: learning new technologies

At the same time, organisational adoption is incomplete: only 34% of development teams have adapted their workflows specifically for AI use.

Changing developer roles and skills

The survey indicates AI will reshape job content rather than eliminate developer roles. Only 6% believe AI could take over most of their current work within five years. By contrast, 67% expect AI agents to become regular members of development teams within five years.

A majority (69%) anticipate an increased role for supervising, coordinating and evaluating outputs from AI systems in their own work. Skills expected to gain importance over the next five years include:

  • 51%: managing AI systems and verifying outputs
  • 51%: critical thinking
  • 47%: system architecture

By comparison, only 15% consider manual coding to remain a defining long-term skill. Employer expectations may shift accordingly: respondents expect curiosity and an experimental mindset (37%) to become far more important than years of experience in a particular programming language (9%).

Adaptation is already underway: 34% said they have had to significantly expand their professional knowledge over the past two years due to the rapid spread of AI.

Education and training lag behind, developers say

Seventy-eight percent of respondents think educational institutions are not keeping pace with AI developments. Seventy-two percent want university computer science curricula to change more rapidly, and 56% believe secondary education should place greater emphasis on AI-related topics.

Despite these concerns, the developer career remains attractive: 68% would recommend the profession to young people, citing competitive salaries and favourable career prospects. Fifty-eight percent also believe AI can make it easier for junior developers to enter the field by enabling them sooner to perform value-creating tasks.

Respondents also flagged new risks: 45% worry about excessive dependence on AI, 41% about a weakening of juniors’ fundamental skills, and 31% about overtrust in AI-generated outputs.

Organisational and educational responses are needed

"AI is fundamentally transforming the developer profession, but tools alone are not enough. At an individual level developers see clear productivity gains; at an organisational level the benefits of AI are realised only if companies rethink the entire development process. Education must also adapt faster so students can enter a labour market where AI is an integral part of everyday work," said Gianni Romolo, Amazon Web Services head of business development for Central and Eastern Europe.

The study ran parallel surveys in Poland and Romania. DevHorizon Hungary 2026 is based on a 500-respondent sample from Hungary and was conducted in line with MRS and ESOMAR guidelines.