I’ve known hundreds of software developers during my career in financial services. Maybe even thousands. What nearly all of them share is the pride and respect they have for the craft. Good code takes time, discipline, and judgment.
So, what happens now that AI can generate working code in seconds and then test it, fix it, and deploy it? Many are asking if the coding profession is on borrowed time.
But that framing misses what’s changing and why.
History tells us that when building becomes easier, the world needs more builders. E-commerce opened new markets, so more people started selling. Cloud computing widened what was possible, so businesses built more software. AI is now bringing that same expansion to software development. It makes code easier to produce and puts more value into the decisions that can’t be automated away. That is the human part.
AI will eliminate the idea that a developer’s value is typing code. The developer role is evolving into something broader and more strategic: software engineering. Great engineers define the problem, design for resilience, and embrace system-level thinking. They guide and validate complex architecture to ensure the product is safe and accurate for customers. High-quality engineers know when a feature is ready and when an AI-generated answer is wrong. They understand when a faster release creates more risk. That judgment is still important.
Together, AI and developers can turn more good ideas into secure, reliable, and valued products. At Synchrony, we’re already seeing this happen. In early pilots with our engineering teams, busywork and rework has shrunk and quality risen, with an estimated 25% faster delivery, 30% fewer defects, 120% more unit tests, and nearly 40% fewer security vulnerabilities.
TRUST REQUIRES GOVERNANCE
AI also relies on other human traits that it doesn’t possess, including the curiosity and discipline to challenge what it produces. Remember, mistakes including security gaps, flawed assumptions, and even data exposure can ship just as fast as clean code. We know we can build it, but can we trust it? We need engineers to test the output and challenge the assumptions. The engineers know when to reject bad code the tool produces. Ultimately, companies can ship code faster, with better outcomes and more time to spend on problems that require human judgment.
That’s why governance can’t be an afterthought. It must be built into how teams use AI, with clear guardrails on approved tools, data use, human review, traceability, quality, and security checks. AI can support work end to end, but accountability stays with people. Even in the age of AI, that is still the only way to create software customers can rely on.
FOR EARLY CAREER TALENT, AI IS AN ACCELERATOR
Those same guardrails and fast feedback loops affect talent too, especially early-career engineers. Some argue AI makes junior hiring unnecessary. That’s not what we find. Early-career engineers often experiment and get creative with these tools quickly, bringing new ideas to the table. Used well, AI is a programming partner that can explain unfamiliar code, flag downstream effects, and help a junior engineer get unstuck faster.
But the benefits are not automatic. Junior talent contributes faster only when companies invest in the fundamentals: a culture of trust and responsibility and real mentorship. AI can accelerate learning, but it can’t replace ownership. We need early career talent to become tomorrow’s senior engineers and leaders, and leadership is built by learning the craft from the ground up.
A WIDER ON-RAMP FOR SOFTWARE CAREERS
AI can also make the profession more accessible for people who did not start on a traditional path. This is personal for me. Growing up, I didn’t think it was possible for me to be a C-suite leader of a financial services company. I loved computer science, but I opted out of AP computer science in high school because I was the only girl in the room. Moments like that shape who sees a future for themselves in the industry.
What changes that experience is faster feedback, safer ways to ask questions, and more support while you build confidence. For many people, the missing piece is confidence, which requires a better way to practice. And leaders still have to do the work around coaching and setting clear expectations.
Universities are responding too, blending AI with a broad range of disciplines from language to critical thinking. What’s emerging is the foundation for modern development that combines the best of human and machine.
The optimist in me sees that there is no set playbook and that creates opportunity to thrive. The people and teams that improve fastest will practice with the tools every day and build habits around review and quality, as well as accountability. Leaders should support these behaviors by investing in them. They can transform workflow and reward quality and judgment over raw volume and output. Developers should reinvent themselves as engineers by building fluency with AI the same way they build fluency with any core tool. They should own the results, using it often and testing what it gives back.
The future will belong to those who can build software people can trust. And that means more software engineers using judgment to solve harder problems.
Carol Juel is executive vice president and CEO of digital platform at Synchrony.
Â
