We’re in the middle of a shift where code is no longer the bottleneck. What matters more and more is what we build, why we build it, and how fast we can adapt it when the world around us changes.
From code producer to problem solver
AI agents and coding tools have already changed our daily work. Many tasks that used to take hours, like setting up boilerplate, writing simpler functions, generating test code, or producing documentation, can now be done in minutes.
But the tools don’t solve the problem for us. They amplify the difference between
- developers who can formulate the problem, understand the domain and make trade-offs
- developers who just write code based on fuzzy wishes
We see value shifting from “how many lines of code can you write” to “how well can you define and solve the right problem together with both humans and AI”.
That means the most important skills going forward are
- deep understanding of the business and domain
- the ability to break down problems into clear goals, steps and architecture
- the courage to ask: should we even build this?
Being able to code is still necessary but it’s no longer the differentiator by itself. What sets a strong consultant apart is the ability to think, assess and take responsibility.
Architecture, quality and responsibility grow in importance
As it becomes easier to generate code, architecture and quality matter more, not less. AI does help us with
- choosing the right boundaries and responsibilities between services
- designing sustainable data flows and integration patterns
- balancing technical decisions against security, regulation, cost and long-term maintenance
If we deploy AI-assisted coding without solid architecture and guardrails, we simply get technical debt at a much higher speed.
As consultants, our responsibility shifts more towards
- defining principles, direction and rules of the game
- establishing patterns, frameworks and platforms where both humans and AI can build safely
- ensuring quality continuously – not just in code review, but throughout the flow from idea to production
We’re moving from individual code hero to builder of systems, platforms and ways of working.
Human in the loop, for real
As AI takes a larger role in development, human in the loop becomes more than a buzzword. We see it at several levels:
Intent and context: AI is completely dependent on good input. Someone has to formulate goals, constraints, non-functional requirements, risks and edge cases. That is a developer’s craft.
Review and interpretation: AI generates suggestions, but takes no responsibility. We do. Real quality is about being able to read, understand and challenge what AI produces, both technically and from a business perspective.
- Ethics and consequences: The systems we build increasingly shape decisions, content, and user experiences. AI can amplify both the good and the bad. Understanding consequences and setting reasonable boundaries is a human responsibility, and a core part of the job.
Developers move closer to decisions, accountability and impact, not further away.
The consultant role: more partner, less resource
From a consulting perspective, the AI shift makes the pure resource logic less relevant. If someone only wants a pair of hands to write code, there will soon be cheaper alternatives.
But when clients genuinely want to understand AI’s potential, build sustainable solutions, and change ways of working and organisations, then they need consultants who act as partners, not just extra keyboard capacity.
Increasingly, our role is to help customers see which problems are AI problems and which are not, combine AI capabilities with existing platforms, data and processes, and set up teams, development flows and governance where AI is a natural, safe part of everyday work.
We become advisors and co-creators of direction, while still being able to dive deep into code and architecture when needed.
That combination – hands-on and strategic – is where a lot of the value sits.
The profession remains but the profile broadens
We don’t buy into the story that developers will be replaced. We see a different evolution:
- Junior developers get leverage from AI much faster, but need guidance in architecture, quality and domain understanding.
- Senior developers spend less time tinkering and more on direction, big-picture design and responsibility.
- The consultant role shifts towards being the link between business, technology and AI capabilities.
There will always be a need for people who care about how things actually work, are willing to dig into the details when it matters and can take responsibility for making solutions usable, secure and sustainable over time.
That’s why we feel fairly confident: software development as a profession is needed, all the time. It’s just that expectations are higher, the context is more complex, and the value of truly good developers is increasing.
And yes, the profession still comes with a price. Not because we type code, but because we take responsibility for systems that actually work, create real value and are possible to live with over time. In a world where AI accelerates everything, that might be more important than ever.
