Future of software engineering

Earlier this year, we shared insights from the Future of Software Engineering retreat convened by Martin Fowler in Utah. With the pace of
change continuing to accelerate, Thoughtworks and Martin recently brought together senior technology leaders in Switzerland to examine how AI is reshaping software engineering and what organizations should do next.

The report distills the retreat’s discussions into five critical themes:

• Verification becomes the new bottleneck: As AI makes code generation faster, the greater challenge is determining whether that code is correct, secure, maintainable, and aligned with business intent.

• Harness engineering emerges as a new discipline: Effective AI-assisted delivery will increasingly depend on the systems, tools, context, and guardrails built around coding agents.

• The apprenticeship model comes under pressure: As AI takes on more entry-level engineering tasks, organisations will need new ways to develop judgment, experience, and technical depth in the next generation of engineers.

• The executive-engineer expectation gap widens: Leaders may expect immediate productivity gains, while engineering teams are confronting the practical complexities of integrating AI safely and effectively.

• Legacy modernisation presents a near-term opportunity: AI may deliver its clearest value by helping organisations understand, test, and modernise complex existing systems rather than simply generate more new code.

Together, these themes point to a fundamental shift: the future of software engineering is not simply about using AI to write more code. It is about rethinking how software is designed, verified, governed and maintained in an increasingly AI-driven world.