AI-driven automation and cloud-native tooling have shifted technology professionals’ attention to speed and efficiency, often sidelining the cultural aspect fundamental to mature DevOps.
In the last two years, “developer productivity” has become a critical metric to measure the effectiveness of technology teams. Platform teams are rolling out internal developer platforms (IDPs), golden paths, and AI pair programmers at a frantic pace. Many of these investments are paying off; for instance, AI-assisted coding can significantly reduce task times in certain situations. In a recent study, a group of JavaScript developers with access to an AI agent completed a programming task 55.8% faster than the group without AI.
But a pattern is emerging: throughput gains stall, incidents creep up, and the same delivery bottlenecks resurface under new toolchains. These are the faults that prevent an organisation from progressing to more mature DevOps.
The root cause for the stalling out of DevOps processes isn’t a lack of automation; it’s a lack of attention to culture. DevOps was never just pipelines. It is a human system of empathy, collaboration, and shared ownership that enables the tools to work.
Tools Accelerate Delivery, but People Sustain It
The latest DORA State of DevOps report shows two trends in DevOps execution. First, platform engineering and AI are reshaping software delivery. Second, the highest performers still rely on human-centred practices, including stable priorities, user-centricity, and a strong developer experience. High-performing teams invest not only in tooling but also in culture and focus.
AI is an Accelerant, Not a Substitute for Culture
Controlled experiments like the one referenced above show that AI can help to speed software development. But GitHub’s research frames the value of using AI in software development in terms of energy, focus, and satisfaction: human outcomes, rather than just straight speed.
And not every study finds that AI speeds software development, especially for complex work. Another study suggests that for ill-defined problems, AI can slow software development by 20% and add rework.
Culture is the Real Multiplier
DevOps is effective because it recognises software delivery as a socio-technical system in several different ways. These include reducing the cognitive load for teams, shared responsibilities, learning from mistakes, and tracking meaningful metrics.
Practical Steps for Stronger DevOps Culture
Prioritise Collaboration Over Silos
DevOps thrives when teams work together, not in isolation. This includes taking steps like encouraging cross-functional standups, so that delivery decisions are transparent to everyone. This strengthens a feeling of shared ownership of outcomes.
Ensure Reliability Guardrails Are Embedded in the Culture
Set clear objectives with a focus on balancing speed and stability. Enforce guardrails with cross-team conversations that involve development, product, engineering, and operations.
Make Empathy a Daily Practice
Empathy is the backbone of DevOps, but it must be modelled and supported throughout the organisation. Leaders can model it by creating safe spaces for feedback, running team health checks, and celebrating the human side of problem-solving. Embedding team members with other teams for short stints can also deepen understanding across teams.
Use AI and Automation as Support, Not Shortcuts.
Introduce AI tools where they make sense. This includes straightforward, well-defined tasks like writing boilerplate, scaffolding tests, or searching across large codebases. But pair AI usage with strong code review practices, knowledge-sharing sessions, and a culture that values thoughtful design. AI should free people for more meaningful collaboration, not replace it.
Manage Cognitive Load With Clarity and Focus
Too many tools and unclear ownership structures overwhelm teams. Leaders should streamline workflows, reduce context-switching, and clearly define team responsibilities. This can range from documenting “how we work” agreements to adopting frameworks for structured alignment.
Invest in Learning Loops
Blameless postmortems, regular retrospectives, and peer-led learning sessions are what turn mistakes into improvements. Leaders should make space for these rituals, protect the time to run them properly, and follow through on action items. The goal is to embed continuous improvement into the organisation’s DNA.
The organisations that sustain high-performance DevOps may look boring from the outside: fewer heroics, more habits. Their platforms feel small and helpful, not sprawling. Their AI usage is targeted. Their dashboards tell stories people can act on. Their leaders protect attention and maintenance budgets as fiercely as feature roadmaps. And their engineers stick around, because the system honours human limits.
A company interested in improving DevOps should definitely retain the accelerants such as platform engineering, golden paths, and AI assistants. But to get compounding returns rather than one-off bumps, it is important to re-centre DevOps on the people doing the work: their cognitive load, their feedback loops, and their trust in each other.
When empathy and ownership lead the way, tools finally serve their true purpose: amplifying human strengths. They transform good teams into great ones not by replacing judgment or collaboration, but by giving people the space to do their best work together. In a culture built on trust, even the most advanced automation becomes less about speed for its own sake and more about enabling sustainable excellence.
