The world of AI is rapidly evolving before our eyes. Some people have the steering wheel firmly in hand, while others are along for the ride.
I’ve never been deeper in the world of AI than recently and it’s pushed me firmly out of my comfort zone.
Building applications doesn’t come naturally to me. Neither does writing code, debugging software or designing technical architectures. A hobby, rather than a profession. Being realistic (or perhaps optimistic), I hope I’m now (at least) halfway through my career, so the chances of me moving into a hands-on technical role become less likely with each passing year. Ain’t gunna happen.
And yet, here I am building an application.
It’s still a few weeks away from going live, but I’ve spent time this year quietly working on a tool that gives Hirefinity’s clients a clearer picture of who is in process, where candidates are in the hiring journey and how much value we’re bringing. CV-to-interview ratios, CV-to-hire ratios, applicant summaries, pipeline visibility and dashboards that bring everything together in one place. Nothing revolutionary, but something useful that I hope our clients will love. More importantly, something I wouldn’t have attempted a few years ago.
The reason I’ve been able to do it is AI.
Using tools such as Codex, I’ve been able to take an idea and turn it into something tangible. The process hasn’t been as simple as typing a prompt and watching an application magically appear, but equally there is no denying that much of what I’ve built would have been beyond me without these tools. They have lowered the barrier to entry so dramatically that someone with no formal software engineering background can now create something that, not so long ago, would have required a development team.
It’s both exciting and slightly unsettling.
It raises an interesting question. Am I driving the tool, or is the tool driving me?
I don’t think I’m a complete passenger. I know what problem I’m trying to solve, what I want the application to do and what a successful outcome will look like. At the same time, I am relying on the tool to generate solutions that I couldn’t have produced on my own. There are parts of the application that I understand deeply and parts where I am placing a significant amount of trust in the output I’m being given. Thankfully, I’ve not been alone on the journey. I’ve had a supremely experienced friend helping me throughout and adding some important guardrails before anyone uses it in anger. He even showed me how to install VScode and get up and running. Gregg Ward, I salute you sir! You’ve remained patience throughout! Even in the face of my complete incomptence!
The more I think about it though, the more I realise this isn’t really a story about me building an application. It’s a story that mirrors what is happening across almost every area of technology. Ok, perhaps not with a novice Brummie trying to build an app, but hear me out…
Each week I speak to dozens of people working in Quality Engineering and leadership positions. They are all on their own AI journey. Some are doing genuinely remarkable things, rethinking how products are built, accelerating delivery, driving efficiency and creating opportunities that simply didn’t exist before. Others are still trying to understand where AI fits into their world and how it can help them become more effective.
What unites both groups is that they are navigating the same challenge.
The real risk isn’t that AI replaces expertise. It’s that we gradually stop developing it.
When a tool can generate code, produce documentation, suggest solutions and answer technical questions in seconds, there is an understandable temptation to accept the output and move on. Most of the time it looks convincing. Much of the time it is convincing. The danger comes when we lose the ability to challenge it. If we stop asking why something works, whether it is the best approach or what assumptions have been made, we risk becoming increasingly dependent on systems that we no longer fully understand.
It’s a major consideration for both those entering the industry and those already on the treadmill, busy delivering while simultaneously working out how and where AI fits into their world.
That’s the difference between driving and being driven.
The people who will thrive in the years ahead won’t necessarily be those who can generate the most output from AI. They’ll be the people who retain ownership of the decision-making. A true human-in-the-loop approach.
From a Quality Engineering perspective, this feels particularly relevant. As AI makes it easier to generate code at scale, the need to validate that code doesn’t disappear; if anything, it increases. While speed may be some companies’ ultimate goal, quality, reliability, performance and confidence still matter. Somebody still needs to ask the difficult questions, and I know that so many of my friends and contacts in the industry are brilliant at doing just that!
So perhaps the most important question any of us can ask isn’t whether we’re using AI. Most of us already are, and those who aren’t probably will be soon enough.
The more interesting question is whether we’re still holding the steering wheel.