Insights from the report: AI in Software Engineering – Making Sense of the Noise, by HI Technology & Innovation.
AI isn’t just changing how software is written, it’s reshaping how engineers are hired, developed, and led. As AI tools become embedded in day-to-day development, long-standing assumptions about technical ability, seniority, and engineering craft are being challenged.
The implications for talent and culture are profound.
Hiring in the Age of AI: Rethinking Technical Assessment
The rise of AI coding assistants has quietly broken one of the most common hiring practices in software engineering: the take-home coding test.
Leaders report that it’s now nearly impossible to know whether a candidate completed an assessment themselves or relied heavily on AI. As a result, traditional methods of evaluating an engineer’s ability to write clean, correct code are becoming increasingly obsolete.
Instead, interviews are starting to shift focus toward how candidates think, not just what they produce.
More effective assessments now emphasise:
- Critiquing and debugging code, including AI-generated output
- Explaining trade-offs and reasoning through decisions
- For senior roles, evaluating architecture, systems thinking, problem decomposition, and design patterns
As AI-generated code becomes the norm, prompting and collaboration with AI itself is emerging as a core skill. Some organisations are already experimenting with interviews that involve pairing candidates with an AI tool, observing how they:
- Guide the model
- Challenge incorrect outputs
- Integrate AI suggestions into a coherent solution
This approach requires clearly defined standards and best practices, but it reflects a deeper truth: the ability to work effectively with AI is becoming part of what it means to be a capable engineer.
Upskilling Without Deskilling: A Delicate Balance
AI presents a powerful opportunity to democratise technical capability. It lowers the barrier to entry for new languages, frameworks, and unfamiliar domains, allowing engineers to move faster and learn in context.
For experienced engineers, AI often acts as a “just-in-time tutor”, removing the need to memorise syntax and freeing up cognitive space for higher-level concerns like architecture and logic.
But this same dynamic introduces risk.
For junior engineers, heavy reliance on AI can create an illusion of competence, especially when using agentic or end-to-end workflows (often referred to as “vibe coding”). Without strong foundational mental models, engineers may ship working code without truly understanding why it works.
To manage this, leaders are beginning to structure AI usage by career stage:
- Senior engineers benefit most from agentic workflows, large-scale refactoring tools, and automation that removes repetitive toil
- Junior engineers may benefit more from AI as:
- An advanced autocompletion tool
- An “instant Stack Overflow”
- A learning or navigation aid
Some organisations are exploring policies where juniors are encouraged to use AI to:
- Explain unfamiliar code
- Generate tests
- Explore alternative approaches
…but are discouraged from using it to generate core business logic in areas they are still learning.
Code reviews may also need to evolve. Rather than reviewing output alone, senior engineers increasingly review the process, asking juniors to explain the why behind AI-generated blocks to ensure genuine understanding.
Ultimately, the goal of an AI-enabled upskilling strategy is not to remove the cognitive load of engineering, but to shift it.
The Senior Engineer’s New Role: From Coder to Conductor
As AI takes on more of the line-by-line coding, the value of senior engineers is being redefined. Speed of execution matters less. Depth of thinking matters more.
Senior engineers are increasingly acting as conductors, orchestrating the work of both humans and AI. Their responsibilities are shifting toward:
- Reviewing: Acting as the final quality gate, with heightened scrutiny of AI-generated code for correctness and architectural integrity
- System design: Defining the patterns, paradigms, and constraints that guide the rest of the team, including AI tools
- Mentoring and coaching: Teaching junior engineers how to use AI responsibly, while ensuring foundational skills continue to develop
This evolution elevates senior talent rather than diminishing it, but only if organisations recognise and support the shift.
Managing Cultural Resistance to AI
Despite the momentum, not every engineer is enthusiastic about AI.
Leaders report pockets of resistance, particularly among experienced developers who feel that:
- AI produces low-quality or brittle code
- It disrupts established workflows
- It removes the enjoyment or craftsmanship from engineering
Overcoming this resistance rarely works through mandates. Instead, successful teams focus on demonstrating value.
Effective strategies include:
- Showcasing clear wins on painful, low-satisfaction work (e.g. legacy code migrations or repetitive refactors)
- Creating open forums where skeptics can voice concerns without judgement
- Framing AI not as a replacement for expertise, but as a way to eliminate toil and create space for more interesting, creative work
When positioned correctly, AI becomes an enabler of better engineering, not a threat to it.
Engineering Culture Still Matters Most
AI is changing what engineers do, how they’re hired, and how they grow. But it doesn’t remove the need for strong engineering culture, it amplifies its importance.
Organisations that approach AI thoughtfully, align it with career development, and invest in leadership and mentorship are seeing AI elevate their teams. Those that treat it as a shortcut risk weakening the very capabilities they depend on.
The message from leaders is consistent: AI changes the tools, not the responsibility.
Want the full picture?
Download the full report: AI in Software Engineering – Making Sense of the Noise
