If developers are using AI daily and saving hours every week, why are business results not improving at the same pace?
This uncomfortable question was at the heart of Professor Burak Turhan’s presentation at the June SW4E AI Working Group meeting. Drawing on an international study of the financial sector, his message was clear: the key challenge is no longer whether organizations use AI, but whether they can institutionalize it in a way that creates measurable business value.
One number captures the dilemma: over 90% of respondents said generative AI had improved their software development processes — yet only about half felt it had met their expectations.
From Experimentation to Strategic Adoption
The research, conducted by the University of Oulu in collaboration with Istanbul Technical University and DefineX, gathered insights from 641 practitioners and 71 executives across banks, fintechs, and insurance companies in the UK, Türkiye, and the Middle East. The survey was conducted in summer 2025 — and as Turhan noted, in a field moving this fast, some patterns may already have shifted. A follow-up study is underway to track how adoption evolves.
The findings show that daily use of generative AI is already mainstream among software professionals — 70% in Türkiye and 51% in the UK and Middle East report using AI tools daily. However, adoption remains concentrated around coding, debugging, code reviews, refactoring, and documentation, while broader use across architecture, requirements engineering, and validation is still emerging.
Developers report significant productivity gains, with most estimating they save one to six hours each week through AI-assisted development. Yet these benefits are largely based on perception rather than measurement. While practitioners see clear improvements in productivity and quality, executives often struggle to identify corresponding improvements in business KPIs.
This is what Turhan described as the “value gap”: organizations believe AI is helping, but lack the frameworks to prove how much value it actually delivers. The study’s conclusion is that AI readiness should not be measured by license counts or usage rates, but through strategic capabilities: governance, process integration, workforce development, and value measurement.
Where Will the Next Senior Developers Come From?
Beyond productivity, the discussion focused heavily on the future of software engineering talent. Survey respondents expect the greatest impact from AI to be felt in entry-level and mid-level development roles. Participants cautioned against interpreting this as the disappearance of software engineering jobs — rather, the profession is evolving.
But this raises a difficult question for business leaders: if entry-level opportunities continue to shrink, where will the next generation of senior experts come from?
As Turhan noted, software engineering and computer science programmes are already facing declining enrolment in many countries, while companies increasingly expect AI-native skills from new graduates. This combination could create long-term talent bottlenecks if organizations focus solely on short-term efficiency gains.
The discussion also highlighted the risks of over-relying on AI-generated code without sufficient governance. Technical debt, maintainability, validation, and accountability remain critical challenges. Producing code faster does not automatically mean delivering better software.
The conclusion was clear: the organizations that gain the most from generative AI will not necessarily be those that use it the most. Success will come to those that build structured processes, invest in talent development, establish governance mechanisms, and create meaningful ways to measure value.
Why SW4E Hosts These Discussions
Bridging research and business practice is exactly what the SW4E ecosystem exists to do. The AI Working Group brings Finnish software-intensive companies and leading researchers to the same table — turning international research findings into practical questions for Finnish businesses: Where does AI create value for us? How do we measure it? How do we develop our people alongside it?
New studies on AI adoption in software engineering are already in the planning, and the working group will continue to follow how the picture develops.
How is your organization measuring the value of AI in software development? Join the conversation.
About the Speaker
Burak Turhan is Professor of Software Engineering at the University of Oulu and a leading researcher in software engineering, AI adoption, and software quality. His research focuses on helping organizations bridge the gap between emerging technologies and sustainable business value.
Read More
- Follow SW4E on LinkedIn
- Public study: https://www.definex.com/wp-content/uploads/2025/10/BEYOND-CODE_GenAIs-Strategic-Impact-on-the-Entire-Software-Development-Lifecycle.pdf
Based on discussions and presentation materials from the SW4E AI Working Group meeting, June 2026.