Industry

AI reshapes software development: efficiency gains and oversight become central

Artificial intelligence is changing software development across the lifecycle, speeding onboarding, automating routine tasks and improving developer experience without lowering quality when proper controls are in place.

AI reshapes software development: efficiency gains and oversight become central

In recent years, artificial intelligence (AI) has driven substantial changes in software development: it has reshaped workflows and affected the daily work and experience of developers. According to Lufthansa Systems Hungária, AI now plays a role at nearly every stage of the software development lifecycle — from clarifying requirements and generating code to testing, debugging and preparing documentation.

Efficiency and developer experience

Internal surveys at the company show that more than 80 percent of developers use some form of AI-based tool on a daily basis. These solutions reduce repetitive tasks, speed up onboarding of new colleagues and shorten the time spent searching for information. Berecz Dániel, Head of Engineering at Lufthansa Systems, says AI particularly helps developers work with fewer interruptions and greater focus, improving productivity and job satisfaction.

Company measurements indicate that increased efficiency does not necessarily come at the cost of quality: with appropriate review processes and conscious use of harness engineering and context engineering methods, professional standards can be maintained.

Supervision and human oversight remain essential

At the same time, Berecz emphasizes that reviewing the large volumes of AI-generated code remains a human responsibility. In some cases this can increase developers’ cognitive load, although in the longer term specialized AI tools for code review are expected to play a greater role.

"Developer experience is no longer a perks issue but a key competitiveness factor. One of AI’s greatest benefits is that it reduces developers’ mental load and cuts the time-consuming interruptions that slow work and harm concentration," Berecz said.

The role of developers is shifting

He warns against the simplistic view that AI will simply replace developers. Instead, the profession’s focus is changing: as manual coding declines in relative importance, the ability to understand complex systems, make decisions, interpret business needs, and enforce security and quality gains prominence. Future developers must not only know how to program but also how to apply AI tools consciously and safely.

"Developers’ tasks will increasingly involve overseeing AI-created solutions, providing the right context, and making professional decisions. Human thinking and systems-level perspective cannot be automated," Berecz added.

Newcomers face particular challenges

AI’s spread is also reshaping the IT labor market, especially affecting entry-level opportunities for junior developers. Many simpler tasks that once served as on-ramps — such as trivial bug fixes, basic testing, or documentation — can now be handled by AI, reducing traditional learning opportunities through hands-on work in enterprise systems.

International data reflect the trend: the Stanford University Human-Centered Artificial Intelligence institute’s 2026 AI Index report found that demand for 22–25-year-old software developers in the United States fell by nearly 20 percent compared to 2024.

Berecz says this can boost short-term efficiency for companies, but raises long-term questions about workforce replenishment. Young professionals can gain an advantage if they learn to use AI-based developer tools early — for example during university through hackathons, practical meetups, conferences and career fairs — which can smooth and accelerate their transition into the job market.

New risks: quality, security and sustainability

AI adoption also brings new risks. Generated code can contain faulty logic, vulnerabilities or unsuitable patterns, so human expertise and expert review remain indispensable for complex systems and strategic decisions. Moreover, running large language models consumes significant energy, so AI’s environmental impact is receiving more attention.

Lufthansa Systems uses AI to reduce environmental impact as well: intelligent route planning, optimizing fuel consumption and reducing on-board food waste are examples where AI both lowers emissions and improves airlines’ operational efficiency and cost management.

Berecz also warns that excessive reliance on AI can hinder development of problem-solving and analytical skills if professionals delegate too many tasks to machines. The industry is working to find methods to mitigate this risk.

"AI will permanently transform software development, but its real value depends on how organizations apply it — with what professional mindset, what controls and to what ends. One of the key tasks in the coming years will be to harness the technology’s efficiency while preserving expertise, quality and long-term sustainability," Berecz concluded.