How AI is reshaping what it means to be hands-on and why that might open new doors in software engineering.
By Annie Vella, Distinguished Engineer, Westpac New Zealand
I was only a few years into my career when my manager suggested I should consider moving into management. “You’ve got great soft skills; you’d make a great team lead!” It sounded like a compliment, and in many ways it was. But I didn’t take it that way. I hadn’t done a four-year degree in computer science and maths to move into management that fast. How could I possibly lead a team of software engineers if I’d barely had a chance to apply and refine my own skills?
This is a fairly common story, especially among women, who are often encouraged into management, coordination, or delivery roles that rely on communication and planning rather than deep technical work. It’s confusing because it sounds like a promotion – more responsibility, better pay – great, right?
The thing is, management requires a completely different set of skills. It’s more of a step into a parallel career track than a step up. Most managers don’t get to actually build software anymore – there’s just no time. Some even feel strongly that managers shouldn’t code, and that if you are, you’re “doing leadership wrong”.
If you got into software engineering to build things yourself, not just influence others, this conversation isn’t comfortable. Sure, part of you feels proud to be seen as capable and trustworthy enough to lead a team. But another part quietly wonders: does this mean my technical skills aren’t as valued as others’?
I held off for ten years before finally accepting that perhaps I’d reached a level where a change was actually a good idea. Building software no longer felt as challenging, so why not try something different – 1:1s, performance reviews, hiring.
Over the years I’ve helped several people navigate this perplexing decision. I remind them that the choice is always theirs, and it’s never a one-time offer. There will be other opportunities, and if they don’t feel ready to swap code for 1:1s yet, they shouldn’t feel pressured to. Growth doesn’t have to follow a single, linear path.
The skills that make a great leader start with great communication. The ability to listen – with empathy for both individuals and organisations – and convey complex, nuanced, or sensitive information in a way others can truly hear is fundamental. Beyond that comes managing context, planning, coordinating work, and anticipating what might come next. These skills have always been valuable, in many aspects of life.
These patterns, where strong communicators are often steered away from hands-on technical paths, reveal something deeper about how we’ve drawn the boundaries between technical and human work, and how those boundaries are beginning to blur.
Enter the age of AI
The impact AI is having on software engineering is significant and undeniable. It’s the topic of many research projects, including my own. There are growing questions about what it even means to be a software engineer when AI coding assistants can code so well.
Sure, AI isn’t perfect – it hallucinates, it goes down rabbit holes and sometimes it can’t easily back out without our help. But let’s be honest, it’s remarkably capable. Only a few years ago, this kind of technology felt like science fiction and now it’s in our IDEs. As a result, the skills software engineers need may not be the same ones we’ve historically prized.
Several schools of thought believe AI is transforming engineers into orchestrators – conductors of a symphony of AI agents with access to well-defined specs, carefully curated context, and just the right tools. To coordinate all this, engineers must be able to communicate intent precisely through well-crafted prompts that limit ambiguity.
Sound familiar?
If orchestration is the new coding
To many, this may no longer sound like software engineering at all. If I’m not writing the code myself, am I really a software engineer? But in a world where the actual writing of code happens through higher-level instructions given to AI, this might well be the new definition of software engineering.
Think about it: there was a time when we hand-wrote assembly code and then along came compilers. Each layer of abstraction has allowed us to build more complex systems, faster. The current shift is similar, even if the technology is far more advanced.
That doesn’t mean the engineering is gone. Quite the opposite. The careful coordination of AI coding agents still demands traditional engineering thinking: security, context management, memory design, tool access and permissions, cost optimisation, guardrails. Knowing which models to use, when, and how, will be as fundamental as knowing which database or cache suits a given use case. The only difference is that these tools will increasingly help build the software with us.
An unexpected shift
If this really is the direction we’re heading, then perhaps those “soft skills” that once saw many of us nudged into management roles will become core skills for future engineers. It’s a change that education and industry will need to navigate together: preparing engineers not just to build with AI, but to think and collaborate differently because of it, and to value design, orchestration, knowledge management and critical thinking as core parts of engineering. Widening the definition of technical work in this way broadens who can contribute meaningfully and what mastery can be.
As this shift unfolds, educators, leaders, and engineers each have a part to play in shaping how we work alongside AI – not just using it, but learning from it, guiding it, and setting the tone for how it fits into our craft.
For educators
- Encourage students to pair deep technical skills with communication, systems thinking, and knowledge management.
- Treat AI literacy as foundational – not a niche skill, but part of every engineer’s toolkit.
For leaders
- Update your definitions of “technical excellence”. Communication, context management, and orchestration are not side skills – they’re central to success in this new era.
- Create equal opportunities for all engineers to explore and experiment with AI tools.
For engineers
- Practise the art of context engineering: structuring specs, prompts, documentation, and system design so others (including AI agents) can understand intent easily.
- Keep deepening your technical intuition and judgement. AI may write the code, but you still need to know what “good” looks like and when something feels off.
A final thought
AI is changing the rules of the game and, in many ways, it may act as an equaliser. Software engineering has always been open to anyone with curiosity and perseverance, and as AI reshapes how we build, the boundaries of technical work are widening to include not just code, but the people and context that bring it to life.
