I've Spent More Than Two Decades Building Software. AI Is Changing the Process, Not the Purpose.

Bindesh Vijayan

I’ve spent more than two decades building software, and one thing has become increasingly clear to me: while technology continues to evolve, the conversations surrounding it rarely do. I’ve watched that happen during the rise of enterprise software, again during the shift to cloud computing while I was at Microsoft, and now with artificial intelligence. Each new wave brings the same prediction: this changes everything. Developers become obsolete. Software will build itself.

One of the defining moments in my career came when I joined Microsoft and had the opportunity to participate in the beta testing of what would eventually become Microsoft Azure. At the time, cloud computing was still a relatively new concept for many organisations, and there were countless predictions about how it would reshape software development.

Many of those predictions were right. Cloud fundamentally changed how we build, deploy, and scale software. But it didn’t change why we build software in the first place.

The technology certainly changes. The purpose rarely does.

Developers didn’t disappear. Their role evolved. Instead of spending time managing infrastructure and servers, they were able to focus more on building products, meaningfully solving user problems, and delivering greater value.

Today, I see AI following a remarkably similar path.

The conversation is understandably centred on what AI can automate, how quickly it can generate code and whether developers will eventually become obsolete.

Personally, I think we’re asking the wrong question. The critical question isn’t whether AI can write code. It’s whether AI helps us build better software. Because after more than two decades building enterprise systems, I’ve come to realise that the hardest part of software development was never writing code in the first place.

The real challenge has always been translating business needs into practical solutions. It’s understanding users whose requirements constantly evolve. It’s balancing technical possibilities with commercial realities. It’s making hundreds of decisions that determine whether a product becomes genuinely useful or simply another piece of software.

Code has always been the output of that thinking. It has never been the thinking itself.

Today, AI is capable of far more than generating boilerplate code. It can build sophisticated implementations, explain unfamiliar codebases, accelerate testing, identify bugs and dramatically reduce the time it takes to move from an idea to working software. The quality of these outputs has improved enormously, and it continues to improve at an extraordinary pace.

But that doesn’t mean every output is the right one. AI is only as effective as the person guiding it. If you don’t understand software architecture, system design or the problem you’re trying to solve, it’s remarkably easy to be misled by code that looks convincing but isn’t appropriate for the task at hand.

In many ways, AI development follows the same principle software engineers have always understood: the quality of the outcome depends on the quality of the inputs. AI can accelerate execution, but it still relies on human judgement to provide direction, context and critical evaluation.

AI isn’t reducing the importance of experienced engineers. It’s changing where their expertise creates the greatest value. For years, developers have spent a significant amount of their time on repetitive implementation work. AI has the potential to reduce much of that effort, allowing engineers to focus on the work that creates the greatest value: designing systems, understanding customer problems, making architectural decisions and building software that genuinely improves the way people work.

At Myndlab, this philosophy shapes everything we build. Our goal has never been to replace developers. It’s to help them move from concept to production more efficiently by removing the friction that slows software development down.

Anyone who has worked with AI coding tools knows that generating code is only part of the challenge. Getting that code to a standard that’s well-structured, production-ready and architecturally sound often requires multiple iterations and constant prompt refinement. We built Myndlab differently by embedding years of software engineering expertise directly into the platform, allowing developers to build with confidence rather than relying on repeated prompt refinement and trial and error to achieve production-quality results. That means spending less time steering the AI, rewriting prompts and validating outputs, and more time building software that is ready for production.

AI should shorten the distance between an idea and a deployable application, not remove the human thinking that makes that application valuable in the first place.

That’s also why I believe the skills that define great software engineers are beginning to evolve.

The conversations I have with engineering teams today are increasingly less about syntax and increasingly more about architecture, system design, model behaviour and understanding how AI can be applied responsibly within real business environments. While prompt engineering has become an important skill today, I don’t believe that should be the future. AI should understand the context provided through specifications and documentation, reducing the need for developers to repeatedly refine prompts just to produce production-quality software. And that is what we have built at Myndlab.

Technical ability will always matter. But as AI takes on more of the implementation, the qualities that increasingly differentiate great engineers are context, judgement and problem solving. The engineers who will thrive won’t simply be those who know how to use the latest AI tools, but those who know which problems are worth solving, how to design systems people can trust and where human judgement remains essential.

Looking back, every major technological shift I’ve experienced has followed a similar pattern:

• Enterprise software changed how organisations operated.
• Cloud changed how software was delivered.
• AI is changing how software is created.
• Each transformation has accelerated the process. None of them have changed the purpose.
• Software has never existed simply to produce code. It exists to solve problems for people.

After decades of building enterprise systems, leading engineering teams and delivering software across different industries, one belief has only become stronger: the tools will continue to evolve. Development will become faster. The distance between an idea and a finished product will continue to shrink. But I don’t believe the purpose of software will ever change.

People don’t need more code. They need better solutions. And if AI allows us to spend more time building those solutions, then I believe it’s one of the most important tools our industry has ever been given. AI is changing the process. Not the purpose.

The ideas I’ve shared here are the same principles that guide how we build at Myndlab. We believe AI should help developers and businesses spend less time on repetitive tasks and more time solving meaningful problems. If you’re exploring how AI can accelerate software development without losing the human judgement that great products depend on, we’d love to continue the conversation.

For further details, please feel free to get in touch with our team.