2025 DORA Report Raises Valid Points On AI Hype Versus Reality
I’ve just been reviewing the 2025 DORA Report, and it’s a firm reminder that we’re living through a real shift in software engineering - but one that’s often sadly being buried under a sea of noise and hype.
AI is everywhere in the headlines, LinkedIn feeds are full of self-proclaimed “AI experts” showing off one-off experiments and quick hacks (never mind shouting about awards for work they often haven’t been directly involved in delivering!).
Sadly, a lot of these experts have never written a line of code pre Chat-GPT, let alone have the ability to validate its output, but we won’t even go there in this article!
The real problem with this level of AI hype? Most of it is built on shaky ground. No solid architecture, no guardrails, no long-term thinking, no ability to scale it beyond an experiment or prototype.
Reality check: this is not innovation - it’s just adding organisational fragility with a shiny new badge.
What the DORA Report really showcases is this:
AI is a force multiplier (for both good and bad).
Yes, it drives speed, throughput, and quality. But it also magnifies weaknesses. If your systems are brittle, AI makes them more brittle.
Engineering is evolving: The role of the engineer is shifting - prompt design, validation, architecture - but it’s still engineering. Not smoke and mirrors.
Systems thinking isn’t optional: Architecture, process, tooling, people, and change. They have to work together. Bolt-ons and shortcuts won’t cut it.
The fundamentals still matter: Testing discipline, modular design, clear ownership, and governance. These aren’t outdated ideas - they’re the only way to stop AI adoption becoming chaos.
Where my AI bets would be placed:
Developer experience as a priority, not an afterthought.
Guardrails and shared patterns, instead of endless “experiments“.
Raising AI fluency across the whole team/organisation, not just a few enthusiasts.
Measuring long-term resilience, not just short-term speed.
The DORA Report effectively cuts through the AI hype, and is aligned with my perspective.
AI is here, but the real wins won’t come from flashy demos or prototypes. They’ll come from teams who invest in foundations, build responsibly, and evolve how they work. People are still involved in these processes, and we still need to design great services and experiences.
Otherwise, it’s just another round of “move fast and break things” - with bigger consequences this time around. The Covid response should have taught us the importance of technical debt.
Executives must understand that there’s no getting away from the hard yards of transformation - don’t be fooled by the AI experts, trust people who are delivery proven because AI is ultimately just another creation tool; otherwise we’ve learned nothing from the technical debt accrued during the previous decade of short term thinking.
Don't be fooled by public sector organisations that are suddenly leading the AI arms race when the reality is they still can’t deliver relatively simple digital experiences for citizens, such as self-service, let alone personalised services. In these cases, using AI is often just putting lipstick on the pig.